1.4K views

Uploaded by Osmar Gabriel Peñaloza

Diseño y analisis de recipientes a presion segun norma

- 05_Process Piping - The Complete Guide to ASME B31.3 - Becht 3rd Ed
- Guide to ASME Code VIII, Division 1, Pressure Vessels
- En13445 vs Asme
- PTB-2_Edtn_2009.pdf
- PTB-7-2014.pdf
- Welding Research Council Bullentin - WRC 537
- Power Piping-The Complete Guide to Asme b31.1 - 2013 - By Charles Becht IV
- 35057
- Pressure Vessel
- Asme Course
- ASME_PTB-5-2013_cont_0101
- Guide to ASME Sec I
- ASME VIII Div 1 Fact Book
- Companion Guide to the ASME Boiler & Pressure Vessel Code Volume 1
- Structural Analysis and Design of Process Equipment
- Asme Viii 2015
- Process Piping_The Complete Guide to ASME B31.3 -Charles Becht IV
- ASME III Design by Analysis - 2006
- ASME Pressure Vessels Basic Calculations
- ASME VIII Div. 1- 1 Internal Pressure

You are on page 1of 229

ANALYSIS OF

ASME BOILER

AND PRESSURE

VESSEL

COMPONENTS

IN THE CREEP

RANGE

by

Maan H. Jawad

Camas, Washington

Robert I. Jetter

Pebble Beach, California

ASME_Jawad_FM.indd

MTC

02/24/2009 06:17AM

2009 by ASME, Three Park Avenue, New York, NY 10016, USA (www.asme.org)

All rights reserved. Printed in the United States of America. Except as permitted

under the United States Copyright Act of 1976, no part of this publication may be

reproduced or distributed in any form or by any means, or stored in a database or

retrieval system, without the prior written permission of the publisher.

INFORMATION CONTAINED IN THIS WORK HAS BEEN OBTAINED

BY THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS FROM

SOURCES BELIEVED TO BE RELIABLE. HOWEVER, NEITHER ASME

NOR ITS AUTHORS OR EDITORS GUARANTEE THE ACCURACY OR

COMPLETENESS OF ANY INFORMATION PUBLISHED IN THIS WORK.

NEITHER ASME NOR ITS AUTHORS AND EDITORS SHALL BE RESPONSIBLE FOR ANY ERRORS, OMISSIONS, OR DAMAGES ARISING OUT OF

THE USE OF THIS INFORMATION. THE WORK IS PUBLISHED WITH THE

UNDERSTANDING THAT ASME AND ITS AUTHORS AND EDITORS ARE

SUPPLYING INFORMATION BUT ARE NOT ATTEMPTING TO RENDER

ENGINEERING OR OTHER PROFESSIONAL SERVICES. IF SUCH ENGINEERING OR PROFESSIONAL SERVICES ARE REQUIRED, THE ASSISTANCE OF AN APPROPRIATE PROFESSIONAL SHOULD BE SOUGHT.

ASME shall not be responsible for statements or opinions advanced in papers or . . .

printed in its publications (B7.1.3). Statement from the Bylaws.

For authorization to photocopy material for internal or personal use under those

circumstances not falling within the fair use provisions of the Copyright Act, contact

the Copyright Clearance Center (CCC), 222 Rosewood Drive, Danvers, MA 01923,

tel: 978-750-8400, www.copyright.com.

Jawad, Maan H.

Design and analysis of ASME boiler and pressure vessel components in the creep

range / by Maan H. Jawad, Robert I. Jetter.

p. cm.

Includes bibliographical references and index.

ISBN 978-0-7918-0284-7

1. Pressure vesselsDesign and construction. 2. Pressure vesselsMaterials.

3. BoilersEquipment and suppliesDesign and construction. 4. MetalsEffect

of temperature on. 5. MetalsCreep. I. Jetter, R. I. II. Title.

TS283.J39 2008

681.76041dc22

2008047421

ASME_Jawad_FM.indd

ii

MTC

02/24/2009 06:17AM

To our wives

Dixie and Betty

ASME_Jawad_FM.indd

iii

MTC

02/24/2009 06:17AM

ASME_Jawad_FM.indd

iv

MTC

02/24/2009 06:17AM

PREFACE

Many structures in chemical plants, reneries, and power generation plants operate at elevated temperatures where creep and rupture are a design consideration. At such elevated temperatures, the

material tends to undergo gradual strain with time, which could eventually lead to failure. Thus, the

design of such components must take into consideration the creep and rupture of the material. In

this book, a brief introduction to the general principles of design at elevated temperatures is given

with extensive references cited for further in-depth understanding of the subject. A key feature of the

book is the use of numerous examples to illustrate the practical application of the design and analysis

methods presented.

The book is divided into seven chapters. The rst chapter is an introduction to various creep topics

such as allowable stresses, creep properties, elastic analog, and reference stress methods, as well as a

few introductory topics needed in various subsequent chapters.

Chapters 2 and 3 cover structural members in the creep range. In Chapter 2, the subject of members

in axial tension is presented. Such members are encountered in pressure vessels as hangers, tray supports, braces, and other miscellaneous components. Chapter 3 covers beams and plates in bending.

Components such as piping loops, tray support beams, internal piping, nozzle covers, and at heads

are included. A brief discussion of the requirements of ANSI B31.1 and B31.3 in the creep region is

given.

Chapters 4 and 5 discuss stress analysis of shells in the creep range. In Chapter 4, various stress

categories are dened and the analysis of various components using load controlled limits of ASME

section III-NH is discussed. Comparisons are also given between the design criteria in VIII-2 and

III-NH and the limitations encountered in VIII-2 when designing in the creep range. Chapter 5 covers the analysis of pressure components using strain and deformation controlled limits. Discussion

includes the requirements and limitations of the A Tests and B Tests outlined in III-NH.

Cyclic loading in the creep-fatigue regime is discussed in Chapter 6. Both repetitive and non-repetitive cycles are presented with some examples illustrating the applicability and intent of III-NH in

non-nuclear applications.

Chapter 7 covers the issues related to buckling of components. Axial members as well as cylindrical

and spherical shells are discussed. Simplied methods are presented for design purposes. The assumptions and limitations required to derive the simplied methods are also given.

The two appendices included in the book are intended as design tools. Appendix A discusses the

derivation of the Bree diagram, used in Chapter 5, and the assumptions made in plotting it. Understanding the derivations will assist the designer in visualizing the applicability of the various regions

in the Bree diagram to various design situations. Appendix B lists some conversion factors for English

and metric units.

ASME_Jawad_FM.indd

MTC

02/24/2009 06:17AM

vi Preface

The design approaches illustrated in this book are based on the experience of the authors over the

past 40 years, with assistance from colleagues. It is the intent of the authors that the methodology

shown in the book will help the engineer accomplish a safe and economical design for boiler and pressure vessel components operating at high temperatures where creep is a consideration.

Maan H. Jawad

Camas, Washington

Robert I. Jetter

Pebble Beach, California

ASME_Jawad_FM.indd

vi

MTC

02/24/2009 06:17AM

ACKNOWLEDGEMENT

This book could not have been written without the help of numerous people and we give our thanks

to all of them. Special thanks are given to Pete Molvie, Bob Schueller, and the late John Fischer

for providing background information on Section I and, to George Antaki, Chuck Becht, and Don

Broekelmann for supplying valuable information on piping codes B31.1 and B31.3.

Our thanks also extend to Don Grifn, Vern Severud, and Doug Marriott for providing insight into

the background of various creep criteria and equations in III-NH and for their guidance.

Special acknowledgement is also given to Craig Boyak for his generous help with various segments

of the book, to Joe Kelchner for providing a substantial number of the gures, to Wayne Mueller and

Jack Anderson for supplying various information regarding the operation of power boilers and heat

recovery steam generators, to Mike Bytnar, Don Chronister, and Ralph Killen for providing various

photographs, to Basil Kattula for checking some of the column buckling equations, and to Ms. Dianne

Morgan of the Camas Public Library for magically producing references and other older publications

obtained from faraway places.

A special thanks is also given to Mary Grace Stefanchick and Tara Smith of ASME for their valuable help and guidance in editing and assembling the book.

vii

ASME_Jawad_FM.indd

vii

MTC

02/24/2009 06:17AM

ASME_Jawad_FM.indd

viii

MTC

02/24/2009 06:17AM

NOTATIONS

Some of the symbols used in this book are dened below

ASME_Jawad_FM.indd

A

A

B

c

C

d

D

D

Dcf

Di

Do

E

EH

EL

Eo

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

Et

f

f

f

F

F

F

G

I

k

k

K

K

Kt

=

=

=

=

=

=

=

=

=

=

=

=

=

=

Ksc

K

Kv

l

L

n

=

=

=

=

=

=

nc

Nd

P

Pa

Pb

Pb

=

=

=

=

=

=

ASME designation for compressive strain in heads and shells

ASME designation for compressive stress in heads and shells

corrosion allowance

at head bending factor in ASME, VIII-1

diameter

Et 3/12(1 - 2)

force-deection matrix of a member

factor to account for the interaction of creep and fatigue damage

inside diameter

outside diameter

modulus of elasticity

modulus of elasticity at hot end of cycle

modulus of elasticity at cold end of cycle

joint efciency factor in ASME, VIII-1, and ligament efciency

in ASME-I

tangent modulus

triaxiality factor

stress reduction factor in pipes

thickness factor for expanded tube ends in ASME, Section I

force in axial members and beams

equivalent peak stress in plates and shells

peak stress in plates and shells

multiaxiality factor

moment of inertia

P/EI

constant

stiffness matrix of an element

plastic shape factor

creep shape factor. Approximate value adopted by the ASME for

a rectangular cross section = (1 + K )/2

stress concentration factor

constant

plastic Poisson ratio adjustment factor

effective length of column

length of member

creep exponent, which is a function of material property and

temperature

number of applied cycles

number of allowable cycles

pressure

ASME allowable external pressure for heads and shells

equivalent primary bending stress

primary bending stress

ix

MTC

02/24/2009 06:17AM

Notations

ASME_Jawad_FM.indd

PL

PL

Pm

Pm

Q

Q

r

Ri

Rm

Ro

Rw

S

Sa

Sm

Sm

Smt

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

Sj

So

=

=

Sr

Sr

Sr

St

Sy

SyH

SyL

t

T

T

X

y

Y

Z

Z

=

=

=

=

=

=

=

=

=

=

=

=

=

=

=

1

Dmax

Dmod

=

=

=

=

=

=

=

c

cH

=

=

=

=

local primary membrane stress

equivalent general primary membrane stress

general primary membrane stress

equivalent secondary stress

secondary stress

radius of gyration = (I/A)0.5

inside radius

mean radius of shell

outside radius

weldment reduction factor based on type of weld rod

allowable stress for I, VIII-1, and VIII-2 construction

alternating cycle stress

(1.5Sm + 0.5St )/3

allowable stress in III-NH

membrane stress. It is the lower value of Sm and St obtained from

III-NH

the initial stress level for cycle type j

Design stress values. The values are taken as equal to Sm except for

a few cases at lower temperatures, where values of Smt at 300,000

hours exceed the Sm values. In those limited cases, So is equal to Smt at

300,000 hours

stress to rupture strength given in Table I-14.6 of III-NH

relaxed stress level at time T adjusted for the multiaxial stress state

relaxed stress level at time T based on a uniaxial relaxation model

time-dependent stress intensity values obtained from III-NH

yield stress

yield stress at the high temperature end of a cycle

yield stress at the low temperature end of a cycle

thickness

time

temperature

primary stress/Sy

temperature coefcient in ASME, Section I

secondary stress/Sy

section modulus

dimensionless effective creep parameter. It represents core

stress values

coefcient of thermal expansion

creep strain

[3(1 - 2 )/Rm2t 2 ]0.25

Ro/Ri

Ri/Ro

maximum equivalent strain range

modied maximum equivalent strain range that accounts for

the effects of local plasticity and creep

total strain range

Poissons ratio

elastic core stress at a cross section

elastic core stress at the high temperature end of a cycle

MTC

02/24/2009 06:17AM

Notations xi

cL

L

r

R

y

1, 2, 3

ASME_Jawad_FM.indd

=

=

=

=

=

=

=

longitudinal stress

radial stress

reference stress

yield stress

circumferential (hoop) stress

principal stresses

xi

MTC

02/24/2009 06:17AM

ABBREVIATIONS FOR

ORGANIZATIONS

AISC

ANSI

API

ASM

ASME

ASTM

BS

EN

MPC

UBC

WRC

ASME_Jawad_FM.indd

American National Standards Institute

American Petroleum Institute

American Society of Metals

American Society of Mechanical Engineers

American Society for Testing and Materials

British Standard

European Standard

Materials Properties Council

Uniform Building Code

Welding Research Council

xii

MTC

02/24/2009 06:17AM

CONTENTS

Preface ..............................................................................................................................................

Notations .......................................................................................................................................... ix

Abbreviations for Organizations ...................................................................................................... xii

Chapter 1

Basic Concepts ..................................................................................................................................

1.1 Introduction .............................................................................................................................

1.2 Creep in Metals ........................................................................................................................

1.2.1 Description and Measurement .....................................................................................

1.2.2 Elevated Temperature Material Behavior .....................................................................

1.2.3 Creep Characteristics ...................................................................................................

1.3 Allowable Stress .......................................................................................................................

1.3.1 ASME B&PV Code ......................................................................................................

1.3.2 European Standard EN 13445 ......................................................................................

1.4 Creep Properties ......................................................................................................................

1.4.1 ASME Code Methodology ...........................................................................................

1.4.2 Larson-Miller Parameter ..............................................................................................

1.4.3 Omega Method .............................................................................................................

1.4.4 Negligible Creep Criteria ..............................................................................................

1.4.5 Environmental Effects ..................................................................................................

1.4.6 Monkman-Grant Strain ................................................................................................

1.5 Required Pressure Retaining Wall Thickness ..........................................................................

1.5.1 Design by Rule .............................................................................................................

1.5.2 Design by Analysis .......................................................................................................

1.5.3 Approximate Methods..................................................................................................

1.6 Effects of Structural Discontinuities and Cyclic Loading ........................................................

1.6.1 Elastic Follow-Up ........................................................................................................

1.6.2 Pressure-Induced Discontinuity Stresses......................................................................

1.6.3 Shakedown and Ratcheting ..........................................................................................

1.6.4 Fatigue and Creep-Fatigue ...........................................................................................

1.7 Buckling and Instability ...........................................................................................................

3

3

3

3

4

7

10

10

12

15

15

15

17

17

19

19

19

19

20

20

25

25

28

29

34

37

Chapter 2

Axially Loaded Members.................................................................................................................. 41

2.1 Introduction ............................................................................................................................. 41

2.2 Design of Structural Components Using ASME Sections I

and VIII-1 as a Guide ........................................................................................................... 45

ASME_Jawad_FM.indd

xiii

MTC

02/24/2009 06:17AM

xiv Contents

2.3

2.4

as a Guide Creep Life and Deformation Limits................................................................... 51

Reference Stress Method ......................................................................................................... 57

Chapter 3

Members in Bending.........................................................................................................................

3.1 Introduction .............................................................................................................................

3.2 Bending of Beams ....................................................................................................................

3.2.1 Rectangular Cross-Sections ..........................................................................................

3.2.2 Circular Cross-Sections ................................................................................................

3.3 Shape Factors ..........................................................................................................................

3.3.1 Rectangular Cross-Sections ..........................................................................................

3.3.2 Circular Cross-Sections ................................................................................................

3.4 Deection of Beams .................................................................................................................

3.5 Piping Analysis ANSI 31.1 and 31.3 ...................................................................................

3.5.1 Introduction .................................................................................................................

3.5.2 Design Categories and Allowable Stresses ...................................................................

3.5.3 Creep Effects ................................................................................................................

3.6 Stress Analysis .........................................................................................................................

3.6.1 Commercial Programs ..................................................................................................

3.7 Reference Stress Method .........................................................................................................

3.8 Circular Plates .........................................................................................................................

61

61

61

62

63

66

66

68

69

72

72

73

75

75

81

81

83

Chapter 4

Analysis of ASME Pressure Vessel Components:

Load-Controlled Limits .................................................................................................................... 87

4.1 Introduction ............................................................................................................................. 87

4.2 Design Thickness ..................................................................................................................... 89

4.2.1 Section I ....................................................................................................................... 90

4.2.2 Section VIII .................................................................................................................. 91

4.3 Stress Categories ...................................................................................................................... 93

4.3.1 Primary Stress .............................................................................................................. 93

4.3.2 Secondary Stress, Q ............................................................................................................. 95

4.3.3 Peak Stress, F .............................................................................................................. 95

4.3.4 Separation of Stresses ................................................................................................... 95

4.3.5 Thermal Stress .............................................................................................................. 99

4.4 Equivalent Stress Limits for Design and Operating Conditions ............................................... 99

4.5 Load-Controlled Limits for Components Operating in the Creep Range ................................105

4.6 Reference Stress Method .........................................................................................................113

4.6.1 Cylindrical Shells ..........................................................................................................114

4.6.2 Spherical Shells ............................................................................................................121

4.7 The Omega Method .................................................................................................................122

Chapter 5

Analysis of Components: Strain- and

Deformation-Controlled Limits ........................................................................................................127

5.1 Introduction .............................................................................................................................127

5.2 Strain- and Deformation-Controlled Limits.............................................................................127

5.3 Elastic Analysis ........................................................................................................................128

5.3.1 Test A-1 ........................................................................................................................128

ASME_Jawad_FM.indd

xiv

MTC

02/24/2009 06:17AM

Contents

xv

5.3.3 Test A-3 ........................................................................................................................130

5.4 Simplied Inelastic Analysis ....................................................................................................137

5.4.1 Tests B-1 and B-2 .........................................................................................................141

5.4.2 Test B-1 ........................................................................................................................141

5.4.3 Test B-2 ........................................................................................................................141

5.4.4 Test B-3 ........................................................................................................................142

Chapter 6

Creep-Fatigue Analysis.....................................................................................................................151

6.1 Introduction .............................................................................................................................151

6.2 Creep-Fatigue Evaluation Using Elastic Analysis ....................................................................151

6.3 Welded Components................................................................................................................174

6.4 Variable Cyclic Loads ..............................................................................................................174

6.5 ASME Code Procedures ..........................................................................................................175

6.6 Equivalent Stress Range Determination ..................................................................................175

6.6.1 Equivalent Strain Range Determination Applicable

to Rotating Principal Strains ........................................................................................175

6.6.2 Equivalent Strain Range Determination Applicable

When Principal Strains Do Not Rotate ........................................................................176

6.6.3 Equivalent Strain Range Determination Acceptable

Alternate When Performing Elastic Analysis................................................................176

Chapter 7

Members in Compression .................................................................................................................183

7.1 Introduction .............................................................................................................................183

7.2 Design of Columns...................................................................................................................183

7.2.1 Columns Operating at Temperatures below the Creep Range ......................................183

7.2.2 Columns Operating at Temperatures in the Creep Range ............................................187

7.3 ASME Design Criteria for Cylindrical Shells under Compression...........................................191

7.3.1 Axial Compression of Cylindrical Shells Operating at

Temperatures below the Creep Range ..........................................................................191

7.3.2 Cylindrical Shells under External Pressure and

Operating at Temperatures below the Creep Range .....................................................192

7.3.3 Cylindrical Shells Subjected to Compressive Stress and

Operating at Temperatures in the Creep Range ...........................................................195

7.4 ASME Design Criteria for Spherical Shells under Compression .............................................198

7.4.1 Spherical Shells under External Pressure and Operating

at Temperatures below the Creep Range ......................................................................198

7.4.2 Spherical Shells under External Pressure and Operating

at Temperatures in the Creep Range ............................................................................199

Appendix A

Background of the Bree Diagram .....................................................................................................201

Appendix B

Conversion Table..............................................................................................................................212

References ........................................................................................................................................213

Index.................................................................................................................................................217

ASME_Jawad_FM.indd

xv

MTC

02/24/2009 06:17AM

ASME_Jawad_FM.indd

xvi

MTC

02/24/2009 06:17AM

ASM_Jawad_Ch01.indd

MTC

02/24/2009 02:18AM

CHAPTER

1

BASIC CONCEPTS

1.1

INTRODUCTION

Many vessels and equipment components encounter elevated temperatures during their operation.

Such exposure to elevated temperature could result in a slow continuous deformation, creep, of the

equipment material under sustained loads. Examples of such equipment include hydrocrackers at reneries, power boiler components at electric generating plants, turbine blades in engines, and components in nuclear plants. The temperature at which creep becomes signicant is a function of material

composition and load magnitude and duration.

Components under loading are usually stressed in tension, compression, bending, torsion, or a

combination of such modes. Most design codes provide allowable stress values at room temperature

or at temperatures well below the creep range, for example, the codes for civil structures such as the

American Institute of Steel Construction and Uniform Building Code. Pressure vessel codes such as

the American Society of Mechanical Engineers Boiler and Pressure Vessel (ASME B&PV) Code,

British, and the European Standard BS EN 13445 contain sections that cover temperatures from the

cryogenic range to much high temperatures where effects of creep are the dominate failure mode. For

temperatures and loading conditions in the creep regime, the designer must rely on either in-house

criteria or use a pressure vessel code that covers the temperature range of interest. Table 1.1 gives a

general perspective on when creep becomes a design consideration for various materials. It is broadly

based on the temperature at which creep properties begin to govern allowable stress values in the

ASME B&PV Code. There may be other specic considerations for a particular design situation, e.g.,

a short duration load at a temperature above the threshold values shown in Table 1.1. These considerations will be discussed later in this chapter in more detail.

It will be assumed in this book that material properties are not degregated due to process conditions. Such degradation can have a signicant effect on creep and rupture properties. Items such as

exfoliation (Thielsch, 1977), hydrogen sulde (Dillon, 2000), and other environment may have great

inuence of the creep rupture of an alloy, and the engineer has to rely on experience and eld data to

supplement theoretical analysis.

One of the concerns to design engineers is the recent increase in allowable stress values in both

Divisions 1 and 2 of Section VIII and their effect on equipment design such as hydrotreaters. Recent

increase in allowable stress reduces the temperature where creep controls and upgrading older equipment based on the newer allowable stress requires knowledge of creep design covered in this book.

1.2

1.2.1

CREEP IN METALS

Description and Measurement

Creep is the continuous, time dependent deformation of a material at a given temperature and

applied load. Although, conceptually, creep will occur at any stress level and temperature if the

3

ASM_Jawad_Ch01.indd

MTC

02/24/2009 02:18AM

Chapter 1

TABLE 1.1

APPROXIMATE TEMPERATURES1 AT WHICH

CREEP BECOMES A DESIGN CONSIDERATION

IN VARIOUS MATERIALS

Temperature

F

Material

Carbon and low alloy steel

Stainless steels

Aluminum alloys

Copper alloys

Nickel alloys

Titanium and zirconium alloys

Lead

700900

370480

8001000

425535

300

150

300

150

9001100

480595

600650

315345

Room temperature

chemistry and failure mode under consideration.

measurements are taken over very long periods, there are practical measures of when creep becomes

signicant for engineering considerations in metallic structures.

Metallurgically, creep is associated with the generation and movement of dislocations, cavities,

grain boundary sliding, and mass transport by diffusion. There are many studies on these phenomena

and there is an extensive literature on the subject. Fortunately for the practicing engineer, a detailed

mastery of the metallurgical aspects of creep is not required to design reliable structures and components at elevated temperature. What is required is a basic understanding of how creep is characterized and how creep behavior is translated into design rules for components operating at elevated

temperatures.

A creep curve at a given temperature is experimentally obtained by loading a specimen at a given

stress level and measuring the strain as a function of time until rupture. Figure 1.1 conceptually shows

a standard creep testing machine. A constant force is applied to the specimen through a lever and

deadweight load. Typically, the test specimen is surrounded by an electrically controlled furnace.

Because creep is highly temperature dependent, considerable care must be taken to ensure that the

specimen temperature is maintained at a constant value, both spatially and temporally.

There are various methods for measuring strain. Figure 1.2 shows one such arrangement suitable

for higher temperatures and longer times, which uses two or three extensometers arranged concentrically around the specimen. Penny and Marriott (1995) have summarized the effects of test variables

on typical test results. They concluded that faulty measurement of mean stress and temperature are the

largest sources of error and that these measurements should be accurate to better than 1% and %,

respectively, to achieve creep strain measurement accuracy to within 10%. For example, it is recommended that, in order to minimize bending effects, tolerances to within 0.002 in. must be achieved in

aligning a -in. diameter specimen.

1.2.2

The distinguishing feature of elevated temperature material behavior is whether signicant creep

effects are present. Consider a uniaxial tensile specimen with a constant applied load at a given temperature. As shown in Fig. 1.3, if the temperature is low enough that there is no signicant creep,

then the stress and strain achieve their maximum values at time t0 and remain constant as long as the

load is maintained. The stress and strain are thus time-independent. However, as shown in Fig. 1.4, if

ASM_Jawad_Ch01.indd

MTC

02/24/2009 02:18AM

Basic Concepts 5

FIG. 1.1

STANDARD CREEP TESTING MACHINE (COURTESY OF ASM)

the test temperature is high enough for signicant creep effects, the strain will increase with time and

eventually, depending on time, temperature, and load, rupture will occur. In the later case, the strain

is time-dependent.

In the previous example, the load was held constant. Now, consider the case with the specimen

stretched to a constant displacement. In this case, as shown in Fig. 1.5, line (a), if the temperature is

low enough that there is no signicant creep, then both the stress and strain will be constant. However,

ASM_Jawad_Ch01.indd

MTC

02/24/2009 02:18AM

Chapter 1

FIG. 1.2

EXTENSOMETER FOR ELEVATED TEMPERATURE CREEP TESTING

(COURTESY OF ASM)

if the temperature is high enough for signicant creep, the stress will relax while the strain is constant

(line b). The behavior illustrated by line (a) is time-independent and by line (b) time-dependent.

Note also the difference in structural response between the constant applied load and the constant

applied displacement. In the rst case, referred to as load-controlled, the stress did not relax and, at

elevated temperature, the strain increased until the specimen ruptured. The membrane stress in a pressurized cylinder is an example of a load-controlled stress. In the second case, referred to as deformationcontrolled, the strain was constant and the stress relaxed without causing rupture. Certain stresses

resulting from the temperature distribution in a structure are an example of deformation-controlled

stresses. Load-controlled stresses can result in failure in one sustained application, whereas failure

due to a deformation-controlled stress usually results from repeated load applications. However, due

to stress and strain redistribution effects (discussed in more detail in subsequent chapters), actual

FIG. 1.3

LOAD CONTROLLED LOADING AT LOW TEMPERATURE

ASM_Jawad_Ch01.indd

MTC

02/24/2009 02:18AM

Basic Concepts 7

FIG. 1.4

LOAD CONTROLLED LOADING AT ELEVATED TEMPERATURE

structures behavior is more complex. For example, if there is elastic follow-up, then the stress relaxation will be slowed down and there will be an increase in strain as shown in Fig. 1.5, lines (c) and

(d), respectively. Thus, elastic follow-up, depending on the magnitude of the effect, can cause deformation-controlled stresses to approach the characteristics of load-controlled stresses. The distinction

between load-controlled and displacement-controlled response and the role of elastic follow-up, or,

more generally, time dependent stress and strain redistribution, is central to the development and

implementation of elevated temperature design criteria.

1.2.3

Creep Characteristics

A representative set of creep curves is shown in Fig. 1.6 for carbon steel. As shown in Fig. 1.7, the

curve is usually divided into three zones. The rst zone is called primary creep and is characterized

by a relatively high initial creep rate that slows to a constant rate. This constant rate characterizes

FIG. 1.5

STRAIN CONTROLLED LOADING AT ELEVATED TEMPERATURE

ASM_Jawad_Ch01.indd

MTC

02/24/2009 02:18AM

Chapter 1

FIG. 1.6

CREEP CURVES FOR CARBON STEEL (HULT, 1966)

the second zone called secondary creep. For many materials, the major portion of the test duration is

spent in secondary creep. The third zone is called tertiary creep and is characterized by an increasing

creep rate that culminates in creep rupture. Although for many materials the major portion of the duration of the test is spent in secondary creep, for some materials for example, certain nickel-based

alloys at very high temperatures primary and secondary creep are virtually negligible and almost the

entire test is in the third stage or tertiary creep zone.

As described more fully in Section 1.4.6, it is sometimes assumed that deformations and stresses

in the primary creep regime do not signicantly contribute to accumulated creep rupture damage.

An interesting application of the above assumption occurs in the assessment of the impact of heat

treatment on structural integrity. For very large components, in particular, the complete time for the

whole heat treating cycle can be quite signicant. Thus, if it were possible to ensure that the heat

treating cycle did not exceed the time duration of primary creep, then one could rationalize that the

FIG. 1.7

CREEP REGIMES STRAIN VS. TIME AT CONSTANT STRESS

ASM_Jawad_Ch01.indd

MTC

02/24/2009 02:18AM

Basic Concepts 9

time spent in heat treatment would not signicantly compromise the functional structural integrity

of the component. Clearly, the key to this approach is to have an estimate of the time to the point at

which primary creep ends and secondary begins. To obtain a general idea of the relevant time duration of primary creep, there is an evaluation by Larke and Parker in a volume edited by Smith and

Nicolson (1971) where they have plotted creep data and analytical correlations for a 0.19% carbon

steel at 842F (450C). In Fig. 1.8, it can be seen that the duration of primary creep depends on stress

level, varying from 300 to about 1200 hours, indicating that a total cycle time of 150200 hours should

be acceptable.

Another means of characterizing creep is to plot isochronous stress-strain curves. Outwardly,

these curves resemble conventional stress-strain curves except that the strain on the abscissa is the

strain that would be developed in a given time by the stress given on the ordinate as shown in Fig. 1.9.

These stress-strain values are usually plotted as a family of curves, each for a constant time as shown in

Fig. 1.10 for 316 stainless steel at 1200F. Although, conceptually, these curves could be directly plotted from data, the curves are usually generated from creep laws, generated from experimental data,

which correlate stress, strain, and time at a constant temperature. These curves can be very useful

in designing at elevated temperature where they can be used similarly to a conventional stress-strain

curve in some situations, i.e., evaluating buckling and instability and as a means of approximating accumulated strain.

Example 1.1

The effective stress in a pressure vessel component is 8000 psi. The material and temperature are

shown in Fig. 1.10. What is the expected design life of the component if:

(a) A strain limit of 0.5% is allowed?

(b) A strain limit of 1.0% is allowed?

FIG. 1.8

MEASURED AND CALCULATED TENSILE CREEP CURVES PRIMARY CREEP

DURATION (SMITH AND NICOLSON, 1971)

ASM_Jawad_Ch01.indd

MTC

02/24/2009 02:18AM

Chapter 1

10

FIG. 1.9

(a) FAMILY OF CREEP CURVES CONVENTIONALLY PLOTTED AS STRAIN VS. TIME AT

CONSTANT STRESS. (b) RESULTANT STRESS-STRAIN CURVES PLOTTED AS

STRESS VS. STRAIN AT CONSTANT TIME

Solution

(a) In Fig. 1.10, the expected life is 15,000 hours.

(b) In Fig. 1.10, the expected life is 65,000 hours.

1.3

1.3.1

ALLOWABLE STRESS

ASME B&PV Code

The ASME B&PV Code lists numerous materials that meet the ASTM as well as other European

and Asian specications. It provides allowable stresses for the various sections of the Code for temperatures below the creep range and at temperatures where creep is signicant. For non-nuclear applications, by far the most common, these allowable stress levels are provided as a function of temperature

in Section II, Part D of the B&PV Code.

For Section I and Section VIII, Div 1 (VIII-1) applications, the allowable stress criteria are given in

Appendix 1 of Part D. The allowable stress at elevated temperature is the lesser of: (1) the allowable

stress given by the criteria based on yield and ultimate strength, (2) 67% of the average stress to cause

rupture in 100,000 hours, (3) 80% of the minimum stress to cause rupture in 100,000 hours, and (4)

100% of the stress to cause a minimum creep rate of 0.01%/1000 hours. Above 1500F, however, the

factor on average stress to rupture is adjusted to provide the same time margin on stress to rupture as

ASM_Jawad_Ch01.indd

10

MTC

02/24/2009 02:18AM

Basic Concepts 11

FIG. 1.10

ISOCHRONOUS STRESS-STRAIN CURVES (ASME, III-NH)

existed at 1500F (815C). Although the allowable stress is a function of the creep rupture strength at

100,000 hours, this is not intended to imply that there is a specied design life for these applications.

There are additional criteria for welded pipe and tube that are 85% of the above values. A very large

number of materials are covered in these tables.

Unlike previous editions, the 2007 edition of Section VIII, Div 2 (VIII-2), covers temperatures in

the creep regime. The time dependent allowable stress criteria for VIII-2 are the same as for VIII-1.

However, because the time independent criteria are less conservative, tensile strength divide by a

factor of 2.4 versus 3.5, the temperature at which the allowable stress is governed by time dependent

properties is lower in VIII-2 than VIII-1.

ASM_Jawad_Ch01.indd

11

MTC

02/24/2009 02:18AM

12

Chapter 1

The allowable stress criteria for components of Class 1 nuclear systems covered by Subsection NH

of Section III (III-NH) of the ASME B&PV Code are different than for non-nuclear components. For

these nuclear components, the allowable stress at operating conditions for a particular material is a

function of the load duration and is the lesser of: (1) the allowable stress for Class 1 nuclear systems

based on the yield and ultimate strength; (2) 67% of the minimum stress to rupture in time, T; (3) 80%

of the minimum stress to cause initiation of third-stage creep in time, T; and (4)100% of the average

stress to cause a total (elastic, plastic, and creep) strain of 1% in time, T. Note that these allowable

stress criteria are more conservative than for non-nuclear systems for the same 100,000-hour reference

time. However, because these allowable stresses apply to operating loads and temperatures (Service

Conditions in Section III terminology) that are, in general, not dened as conservatively as the Design

Conditions to which the allowable stresses apply for non-nuclear applications. There are also additional criteria for allowable stresses at welds and their heat affected zone. All these allowable stresses

are given in III-NH for a quite limited number of materials.

The allowable stresses for Class 2 and 3 elevated temperature nuclear systems are in general similar

to those for non-nuclear systems and are provided in Code Case N-253.

Subsection NB (III-NB) covers Class 1 nuclear components in the temperature range where creep

effects do not need to be considered. Specically, III-NB is limited to temperatures for which applicable allowable stress values are provided in Section II, Part D. These temperature limits are 700F

(370C) for ferritic steels and 800F (425C) for austenitic steels and nickel-based alloys.

Unlike Section I and VIII-1 components, the design procedures for nuclear components, particularly Class 1 components, are signicantly different at elevated temperatures as compared to the requirements for nuclear components below the creep regime. This is due in part to the time dependence

of allowable stresses, but, more signicantly, due to the inuence of creep on cyclic life. As compared

to Section I and VIII-1, components, III-NH explicitly considers cyclic failure modes at elevated

temperature, whereas Sections I and VIII-1 do not. Section VIII-3 does address cyclic failure modes

below the creep range.

Section VIII-2 addresses cyclic failure modes and, as previously noted, currently covers temperatures in the creep regime above the previous limits of 700F (370C) and 800F (425C) for ferritic

and austenitic materials, respectively. Section VIII-2 also requires either meeting the requirements

for exemption from fatigue analysis, or, if that requirement is not satised, meeting the requirements

for fatigue analysis. However, above the 700/800F (370/425C) limit, the only available option is to

satisfy the exemption from fatigue analysis requirements because the fatigue curves required for a full

fatigue analysis are limited to 700F and 800F (370C and 425C).

1.3.2

EN 13445 applies to unred pressure vessels It is analogous to VIII-1 and -2 in that it covers both

Design by Formula (DBF) similarly to Div 1 and Design by Analysis (DBA) similarly to Div 2. It

is unlike Section VIII in several important respects. First, the EN 13445 allowable stresses are time

dependent, analogous to what is done in Subsection NH. They are also a function of whether there

is in-service monitoring of compliance with design conditions. Provisions are also made for weld

strength reduction factors, analogous to Subsection NH. Unlike the BDF rules in Section VIII, Div

1, BDF rules in the EN code are only applicable when the number of full pressure cycles is limited

to 500.

The basic allowable stress parameters in EN 13445 in the creep range are the mean creep rupture

strength in time, T, and the mean stress to cause a creep strain of 1% in time, T. For DBF rules, the

safety factor applied to the mean creep rupture stress is 1/1.5 if there is no in-service monitoring, and

1/1.25 if there is. There is no safety factor on the 1% strain criteria. If there is in-service monitoring

then the strain limit does not apply, but strain monitoring is required. Thus, for a design life of 100,000

hours in the EN code without in-service monitoring, the base metal design allowable stress will be the

same as in VIII-1 when the allowable stresses are governed by creep rupture strength (remembering

ASM_Jawad_Ch01.indd

12

MTC

02/24/2009 02:18AM

Basic Concepts 13

that VIII-1 allowable stresses are based on 100,000-hour properties even though there is no specied

design life in VIII-1).

There are two DBA methodologies dened in the EN 13445, the Direct Route and the Method

based on stress categories. Conceptually, the stress category methodology is similar to the methodology dened in Section VIII-2 and III-NB for temperatures below the creep range and in III-NH for

elevated temperatures; however, there are many differences in the details of their application. The

basic allowable stresses for the stress category DBA methodology are the same as for the DBF rules

and are dependent on whether there is in-service monitoring. The Direct Route is based on limit

analysis and reference stress concepts. It is quite complex. Indeed, there is a warning in the introduction cautioning that, Due to the advanced methods applied, until sufcient in-house experience can

be demonstrated, the involvement of an independent body, appropriately qualied in the eld of DBA,

in the assessment of the design (calculations). On that basis, a detailed discussion of the Direct

Route DBA rules in EN 13445 will be considered beyond the scope of this presentation; however,

there is a further discussion of the reference stress concept in Section 1.5.3.2.

FIG. 1.11

RUPTURE STRENGTH (JAWAD AND FARR, 1989)

ASM_Jawad_Ch01.indd

13

MTC

02/24/2009 02:18AM

14 Chapter 1

Example 1.2

Figure 1.11 shows a representative plot of creep rupture data with extrapolation to 100,000 hours.

Figure 1.12 shows a plot of creep strength (minimum creep rate) for the same material. Curve tting

procedures are usually used for the extrapolation. Based on the creep properties shown in Figs. 1.11

and 1.12, calculate the allowable stress at 1200F for an VIII-1 application. (Note: This is for a nonASME Code application. The Code has published values for Code applications.) Compare the results

to the allowable stress for an EN 13445 application with 100,000 hours design life and with in-service

monitoring. (Note: Under EN 13445, allowable stress values are established by the user based on

published properties as described therein.)

Solution

In Fig. 1.11, the average stress to rupture in 100,000 hours is 22 ksi and the allowable stress for

VIII-1 based on average creep rupture is 22 (0.67) = 14.7 ksi. Assuming a minimum value of creep

rupture based on a 20% scatter band gives a minimum creep rupture strength of 17.6 ksi, and an

allowable stress based on minimum creep rupture of 17.6 (0.8) = 14.1 ksi. In Fig. 1.12, the stress

FIG. 1.12

CREEP STRENGTH (JAWAD AND FARR, 1989)

ASM_Jawad_Ch01.indd

14

MTC

02/24/2009 02:18AM

Basic Concepts 15

for a minimum creep rate of 0.01% in 1000 hours is 15 ksi, which gives an allowable stress of 15 ksi.

Therefore, the applicable stress for a Section VIII-1 application is 14.1 ksi governed by the minimum

creep rupture strength.

For EN 13445 applications with in-service monitoring, the safety factor is 1.25 on mean creep

rupture strength (assumed equal to average strength plotted in Fig. 1.11), so the allowable stress for

100,000 hours design life is 22/(1.25) = 17.6 ksi.

1.4

1.4.1

CREEP PROPERTIES

ASME Code Methodology

One of the issues facing the designer of elevated temperature components is how to extrapolate

limited time duration test data to the service lives representative of most design applications or, in

the case of developing allowable stresses for ASME Code applications, 100,000 hours. For the development of ASME Code elevated temperature allowable stresses for non-nuclear applications, the

method is described in Chapter 3 (Basis for Tensile and Yield Strength Values) of the Companion

Guide to the ASME Boiler & Pressure Vessel Code (Jetter, 2002). Quoting from that source:

At the elevated temperature range in which the tensile properties become time-dependent, the

data is analyzed to determine the stress to cause a secondary creep rate of 0.01% in 1,000 hours and

the stress needed to produce rupture in 100,000 hours. This data must be from material that is representative of the product specication, requirements for melting practice, chemical composition,

heat treatment, and product form. The data is plotted on log-log coordinates at various temperatures. The 0.01%/1,000 hour creep stress and the 100,000 hour rupture stress are determined from

such curves by extrapolation at the various temperatures of interest. The values are then plotted on

semi log coordinates to show the variation with temperature. The minimum trend curve denes the

lower bound for 95% of the data.

As part of the ASME Code methodology, data for development of allowable stress values is required for long times, usually at least 10,000 hours for some data, and at temperatures above the range

of interest, usually 100F (40C) higher. Considerable judgment is exercised in the development of

Code allowable stress values and the use of these values is required for Code-stamped construction.

As a corollary, if the material of interest is not listed in the Code for the applicable type of construction, or at the desired temperature, then it is not possible to qualify the component for a Code stamp.

The designer may, however, use this method for non-Code applications.

A somewhat different approach is taken in EN 13445. There, the mean creep rupture strength and

mean stress for a 1% strain limit are listed in referenced standards for approved materials for various

times and temperatures. The requirements for extrapolation or interpolation to other conditions are

dened and the safety factors to be applied are dened as a function of the application as described

above.

1.4.2

Larson-Miller Parameter

As might be expected, there are numerous methods (Conway, 1969) for extrapolation of creep data;

the ASME procedure described above is the one used for establishment of allowable stresses shown

in Section II, Part D. Generally, the use of other extrapolation techniques would only be required for

non-coded construction or for evaluation of failure modes beyond the scope of the applicable code.

Penny and Marriott (1995) provide an extensive assessment of various extrapolation techniques, including the widely used Larson-Miller parameter, which they characterize as simple and convenient

but not particularly accurate.

ASM_Jawad_Ch01.indd

15

MTC

02/24/2009 02:18AM

16 Chapter 1

The starting point for development of the Larson-Miller (Grant, 1965) parameter is to assume that

creep is a rate process governed by the Arrhenius equation

de c /dT = Ae (-Q/RT )

(1.1)

where

A = constant

Q = activation energy for the creep process, assumed a function of stress only

R = universal gas constant

T = absolute temperature (460+F)

T = time

Noting that from the Monkman-Grant relationship the time to rupture, Tr, times the minimum

creep rate can be assumed to be constant, Eq. (1.1) can be rewritten as

ATr e (-Q/RT ) = constant

Taking logarithms of each side, the Larson-Miller parameter, PLM, can be expressed as

(1.2)

T C log 10T

P LM

where PLM is a function of stress and independent of temperature. It was also assumed that C is independent of both stress and temperature and is a function of material only. Experimental data shows

that the range of C for various materials is between 15 and 27. Most steels have an A value of 20.

Hence, Eq. (1.2) can be expressed as

PLM

460 qF20 log10T

(1.3)

The Larson-Miller parameter is also used to correlate creep data using specic values of C that are

material and temperature range dependent, thus minimizing some of the uncertainties.

Another important use of the Larson-Miller parameter is determination of equivalent time at temperature as shown by the following examples.

Example 1.3

A pressure vessel component was designed at 1200F with a life expectancy of 100,000 hours. What

is the expected life if the design were lowered to 1175F?

Solution

The Larson-Miller parameter for the original design condition is obtained from Eq. (1.3) as

PLM

41500

41500

or

log10T

T

ASM_Jawad_Ch01.indd

16

538

241100 hours

MTC

02/24/2009 02:18AM

Basic Concepts 17

which indicates a 2.4-fold increase in the life of the component when the temperature drops 25F.

Example 1.4

A pressure vessel shell is constructed of 2.25Cr-1Mo steel. The thickness is 4 in. and requires a post

weld heat treating at 1300F for 4 hours. The fabricator requires two separate post weld heat treats (8

hours) and the user needs three more post weld heat treats (12 hours) for future repair. Hence, a total

of 20 hours are needed. The material supplier furnishes the steel plates with material properties guaranteed for a minimum of 20 hours of post weld heat treating. During the manufacturers second post

weld heat treat, the temperature spiked to 1325F for 2 hours. How many hours are left for the user?

Solution

Calculate PLM from Eq. (1.3) for 2 hours at 1325F.

P LM

36237

Substitute back into Eq. (1.3) to calculate the equivalent time for 1300F.

36237

T

39 hours

Thus, the fabricator used a total of 4.0 + (2.0 + 3.9) = 9.9 hours.

Available hours for the user = 20 - 9.9 = 10.1 hours. This corresponds to two full post weld heat

treats plus one partial post weld heat treat.

1.4.3

Omega Method

The Omega method (Prager, 2000) is based on a different model for creep behavior than that described above for the Larson-Miller parameter. Originally developed to address the issue of determining

the accumulated damage, and thus the remaining life of service-exposed equipment, the Omega method

is based on the observation that, at design stress levels, both the primary creep and secondary creep

phases are of relatively short duration with small strain accumulation and that most of the component

life is spent in the third stage, where the strain rate is increasing with time and accumulated strain. In

the Omega method, the creep strain rate is accelerated in accordance with the following relationship:

lndH dT

lndH 0 dT : p H

(1.4)

where

(d/dT) and (d0/dT) = current and initial strain rates, respectively

Wp = Omega parameter

= current strain level

From this relationship, various parameters relating to accumulated damage and remaining life may

be developed. The Omega method has been incorporated into API 579 for remaining life assessments.

An example is provided in Chapter 4.

1.4.4

Another issue of interest is the temperature at which creep becomes signicant. To answer this

quantitatively, the key point is, signicant compared to what? There are no single, rigorous criteria

for assessing when creep effects are negligible. However, in each of the design codes of interest, the

criteria for negligible creep applicable to that particular design code are dened.

ASM_Jawad_Ch01.indd

17

MTC

02/24/2009 02:18AM

18 Chapter 1

For Sections I and VIII-1, the comparison is between the results provided by the allowable stress

criteria based on short-time tensile tests without creep and long-term tests with creep. When the allowable stress as a function of temperature is governed by creep properties, the stress value is italicized

in Section II, Part D, Table 1. However, in this case, even though the allowable stress is governed by

creep properties, the design evaluation procedures do not change.

The situation is different with Section III-NH. In NH, there are two sets of allowable stresses for

primary (load-controlled) stresses to be used in the evaluation of Service Conditions. One set, Sm, is

time-independent and a function of short-time tensile tests. The other set, St, is time-dependent and

a function of creep. As will be discussed in more detail later, the design rules for time-independent

and time-dependent allowable stress levels are different. However, it is in the rules for displacementcontrolled stresses, such as thermally induced stresses, that the criteria for negligible creep are the

most restrictive.

The NH criteria for negligible creep for displacement-controlled stresses are based on the idea that,

under maximum stress conditions, creep effects should not compromise the design rules for strain

limits or creep-fatigue damage. The key consideration from that perspective is that the actual stress in

a localized area can be much greater due to discontinuities, stress concentrations, and thermal stresses

than the wall-averaged primary stresses in equilibrium with external loads. Basically, the magnitude of

the localized stress will be limited by the materials actual yield stress because it is at this stress level

that the material will deform to accommodate higher stresses due to structural discontinuities or thermal gradients. Thus, the objective of the negligible creep criteria for localized stresses is to ensure that

the damage due to the effects of creep at the materials yield strength will not signicantly impact the

design rules for the failure mode of concern. For example, there are two resulting criteria, one based

on negligible creep damage and the other based on negligible strain. For negligible creep rupture damage, the III-NH criteria are given by

Ti Tid d

01

(1.5)

where

Ti = the time duration at high temperature

Tid = the allowable time duration at a stress level of 1.5 times the yield strength, Sy

For negligible strain, the criteria are given by

Hi d

02

(1.6)

where

i = the creep strain at a stress of 1.25 times yield strength, Sy

In Code Case N-253, which provides elevated temperature design rules for Class 2 and 3 nuclear

components, Appendix E contains a gure that shows time temperature limits below which creep

effects need not be considered in evaluating deformation-controlled limits. These curves are lower,

smoothed versions of the Subsection NH criteria for negligible creep for a limited number of materials: cast and wrought 304 and 316 stainless steel, nickel-based Alloy 800H, low alloy steel, and carbon

steel. The advantage of these curves is that no computations are required.

The French code for elevated temperature nuclear components, RCC-MR, also provides criteria for

negligible creep, which is somewhat different than that provided in Subsection NH. The procedures are

more involved than those in Subsection NH, but the resulting values for long-term service are similar to

the temperature limits of Subsection NB, 700F for ferritic and 800F for austenitic and nickel-based

alloys. For 316L(N) stainless steel, whose creep properties are fairly close to 316SS, the time-temperature

limit curve is generally in agreement with the curve shown in Code Case N-253 for 316SS.

ASM_Jawad_Ch01.indd

18

MTC

02/24/2009 02:18AM

Basic Concepts 19

1.4.5

Environmental Effects

As stated in its Foreword, the ASME B&PV Code does not specically address environmental effects. However, non-mandatory general guidance is provided in several Sections. Section II, Part D,

Appendix A provides guidance on metallurgical effects including a number of references on corrosion

and stress corrosion cracking. Section VIII, Div 1, Appendix E contains suggested good practice for

determining corrosion allowances, which are the responsibility of the user to specify based on the

equipments intended service. It is noted that the corrosion allowance is in addition to the minimum

required thickness. Section III, Appendix W, has a comprehensive discussion of environmental effects. Included for each phenomenon is a discussion of the mechanism, materials, design, mitigating

actions, and references.

In the context of elevated temperature applications, the designer should be particularly aware of

environments that can reduce a materials creep rupture life and/or ductility. For example, it has been

shown that short-term exposure to oxygen at temperatures exceeding 1650F (900C) could lead to

embrittlement at intermediate temperatures of 13001500F (705C to 815C), which was attributed

to intergranular diffusion of oxygen. Hydrogen, chlorine, and sulfur may also cause embrittlement due

to penetration. Sulfur is of particular concern because it diffuses more rapidly and embrittles more

severely than oxygen.

1.4.6

Monkman-Grant Strain

Another parameter of interest is the strain computed by multiplying the time to rupture by the secondary creep rate. This strain parameter, shown diagrammatically in Fig. 1.7, is sometimes known as the

Monkman-Grant strain. As discussed by Penny and Marriott (1995), this computed strain has been

shown to be useful in correlating rupture under variable loading conditions. A corollary of this approach

is that it implies that the primary creep strain may be disregarded in assessing damage accumulation.

It has also been suggested that a relevant measure of creep ductility for the application of reference

stress methods (Section 1.5.3.2) in the presence of local stress discontinuities is for the material of

interest to show a ratio of total strain at failure to the Monkman-Grant strain of at least 5:1.

There are basically two approaches in general use in design for determining the wall thickness required to resist internal pressure and applied external loads. The rst is usually referred to as Design

by Rule or Design by Formula (DBF in the European Standard terminology) and the second is Design

by Analysis (DBA). As an alternative to DBA, there are other approaches based on experimental

methods; however, those methods are generally not applicable in the creep regime. In addition to the

above approaches, there are many pressure-retaining components that have standardized allowable

pressure ratings as a function of design temperature. Typically, these include anges, piping components, and valve bodies. In general, these pressure/temperature ratings do not include the effects

of loadings other than internal pressure. The following discussion will provide an overview of these

methodologies; the specic requirements for their implementation will be discussed in later chapters.

1.5.1

Design by Rule

In this approach, formulas are provided for the required thickness as a function of the design pressure, allowable stress, and applicable parameters dening the geometry of interest. Numerous diagrams are provided to dene the requirements for specic congurations, for example, reinforcement

of openings, head-to-cylinder joints, and weldments. This is the approach used, for example, in Section

VIII, Div 1 Unred Pressure Vessels, and Section I Power Boilers of the ASME B&PV Code.

ASM_Jawad_Ch01.indd

19

MTC

02/24/2009 02:18AM

20

Chapter 1

1.5.2

Design by Analysis

In the DBA approach, stress levels are determined at various critical locations in the structure

and compared to allowable stress levels, which are a function of the applied loading conditions and

failure mode under consideration. The most commonly used methodology, particularly at elevated

temperatures, is based on elastically calculated stresses, which are sequentially categorized based on

the relevant failure mode. Primary stresses (those that normally determine wall thickness) are rst

determined by separating the structure into simpler segments (free bodies) in equilibrium with external loads. Next, secondary and peak stresses (which in combination with primary stresses normally

determine cyclic life) are determined from stresses at structural discontinuities and induced thermal

stresses. Different allowable stresses are assigned to the different stress categories based on the failure

mode of concern.

1.5.3

Approximate Methods

There is another category of Design-by-Analysis methodologies that are approximate in the sense

that they approximate the true time-dependent stress and strain history in a component. In fact,

considering the variations in creep behavior and difculties encountered in dening comprehensive

models of material behavior, they can be quite useful under appropriate circumstances. Two main approaches will be described. The rst is the elastic analog or stationary creep solution and the second

is the reference stress approach, which is somewhat analogous to limit analysis.

1.5.3.1 Stationary Creep Elastic Analog. Subject to certain restrictions on representation of

creep behavior, a structure subjected to a constant load will reach a condition where the stress distribution does not change with time, thus the term stationary creep. The fundamental restriction on

material representation is that the creep strain is the product of independent functions of stress and

time. Conceptually, stationary creep is valid when the strains and strain rates due to creep are large

compared to elastic strains and strain rates.

If the structure is statically determinate throughout, then the initial stress distribution will not

change with time, subject to the applicability of small displacement theory that applies to the large

majority of practical design problems. Examples would be a single bar with a constant tension load

and the stresses in the wall of a thin-walled cylinder, remote from discontinuities, subjected to a constant internal pressure.

However, it is with indeterminate structures that the stationary creep concept is of most value. In

a structure with redundant load paths or subject to local redistribution, i.e., a beam in bending, it has

been shown that the stress redistribution will take place relatively quickly; on the order of the time it

takes for the creep strain to equal twice the initial elastic strain. For a set of variables representative of

pressure vessels in current use, Penny and Marriott calculated an effective redistribution time of about

100 hours. Although this would be a long time if the vessel were subject to signicant daily cycles, it

is a short time compared to times of extended operation.

A number of investigators have shown that because the stress distribution in stationary creep does

not vary with time, and thus corresponding creep rates are constant, the stationary creep stress distribution is analogous to the non-linear elastic stress distribution, and solutions to the creep problem

can be obtained from solutions to the non-linear elastic stress distribution problem. This is usually

referred to as the elastic analog. Although the elastic analog has been shown as valid in more general terms, a more convenient representation is based on the analogy between a simple power law

representation of steady, secondary creep in which primary creep is considered negligible

dH dT

k Vn

(1.7)

ASM_Jawad_Ch01.indd

20

MTC

02/24/2009 02:18AM

Basic Concepts 21

k 'TV n

(1.8)

H

K Vn

(1.9)

An example of stationary creep solutions for various values of the power law exponent, n, is shown

in Fig. 1.13. This is the non-dimensional stationary creep solution for a beam in bending with a constant applied moment. Note that for n = 1 the stress distribution is elastic, and for n the distribution corresponds to that for the assumption of ideal plasticity. All the distributions pass through a

point partway through the wall referred to as a skeletal point. The reduction in steady creep stress

as compared to the initial elastic distribution is the basis for the reduction of the elastically calculated

bending stress by a section factor when comparing calculated stresses to allowable stress levels in

Subsection NH and other design criteria.

1.5.3.2 Reference Stress. The initial idea of the reference stress was that the creep behavior of a

structure could be evaluated by use of the data from a single creep test at its reference stress. Initially

applied to problems of creep deformation, there were a number of analytical solutions developed for

specic geometries. However, Sim (1968), noting that reference stress is independent of creep exponent

FIG. 1.13

STEADY-STATE CREEP STRESS DISTRIBUTION ACROSS A RECTANGULAR BEAM IN

PURE BENDING AND HAVING A STEADY-STATE CREEP LAW OF THE FORM C = A

(OAK RIDGE NATIONAL LABORATORY)

ASM_Jawad_Ch01.indd

21

MTC

02/24/2009 02:18AM

22

Chapter 1

and also that the solution for an innite creep exponent is analogous to the limit solution corresponding to ideal plasticity, proposed that the reference stress could be conservatively obtained from

VR

PPL V y

(1.10)

where

P = load on the structure

PL = limit value of the load

R = reference stress

y = the yield stress

There have been numerous comparisons between the results of this approach and experimental and

rigorous analyses of the same component or test article. In general, the results are quite favorable. Although the reference stress approach has not been incorporated in the ASME B&PV Code, it has been

used in the British elevated temperature design code for nuclear systems, R5, and in the recent European Standard EN 13445. However, the British standard recommends an adjusted reference stress for

design given by a factor of 1.2 times the reference stress from the Sims relationship given above. Also,

as previously noted, the EN standard cautions against the use of the reference stress method by those

not familiar with its application.

Part of the reason for this concern is inherent in the basis for both limit loads and reference stress

determination. Both are based on structural instability considerations and not local damage. As such,

there is an inherent requirement that the material under consideration be sufciently ductile. This is

easier to achieve at temperatures below the creep range. Within the creep range, ductility decreases,

particularly at the lower stress levels associated with design conditions. There have been some studies to more specically identify creep ductility requirements, but current thinking would put it in the

range of 5% to 10% for balanced structures without extreme strain concentrating mechanisms.

The following example highlights the differences between an elastically calculated stress distribution, a steady stationary creep stress distribution, and the reference stress distribution.

Example 1.5

Consider the two-bar model shown in Fig. 1.14. As explained in Chapter 2, this is actually representative of the way in which cyclone separators are sometimes hung from vessels. For this example, the

two, parallel, uniaxial bars are of equal area, A, unequal lengths L1 and L2, and attached to a rigid boss

constrained to move in the vertical direction only. The assembly is loaded with a constant force F.

Compare the (1) initial elastic stress, (2) stationary stress, and (3) reference stress in each bar. Assume

that creep is modeled with a power law with exponent n = 3. Consider two cases. In the rst case,

L1 = L2/8 and in the second case, L1 = L2/2.

Solution

(1) Elastic analysis

The initial elastic distribution can be expressed as

H 1 0

V 1 0E and H 2 0

V 2 0E

(a)

where

1(0), 1(0) and 2(0), 2(0) = initial strain and stress in bars 1 and 2, respectively

E = modulus of elasticity

From equilibrium

F

ASM_Jawad_Ch01.indd

22

A V 1 V 2

MTC

(b)

02/24/2009 02:18AM

Basic Concepts 23

FIG. 1.14

TWO-BAR MODEL WITH CONSTANT LOAD

L1H 1 0

L2H 2 0

(c)

L1V 1

L2 V 2 or V 2

L1 L2 V 1

(d)

F

A >V 1 L1 L2 V 1@

(e)

From which, solving for the initial elastic stress distribution gives

V 1 0

>L2 L1 L2 @FA

(f )

V 2 0

>L1 L1 L2 @FA

(g)

The solution for the stationary stress distribution in each bar proceeds in a similar fashion.

The equilibrium equation (a) remains the same.

The strain rate at any time can be expressed as the sum of elastic and creep strain rates

ASM_Jawad_Ch01.indd

23

MTC

02/24/2009 02:18AM

24 Chapter 1

dH 1 dT

k V2n dV 2 dT E

(h)

L1dH 1 dT

(i)

L2dH 2 dT

Using the above relationships, an equation can be developed involving functions of 1; but, as noted

by Kraus (1980), it cannot be solved in closed form. However, we are interested in the stationary creep

solution, where d/dT 0 and (d/dT )/E << K n. Thus, Eq. (h) in the stationary creep regime becomes

dH 1 fdT

k V 1 f n and d H 2 fdT

k V 2f n

( j)

Substituting Eq. ( j) into Eq. (i) provides the following relationship for 2()

V 2 f

1n

1n

L1 L 2 V 1 f

(k)

From the above and the equilibrium equation (b), the following expressions for stationary stress

distribution may be obtained:

1n

1n

1n

V 1 f

>L2 L1 L2 @FA

V 2 f

>L1 L1 L2 @FA

1n

1n

1 n

(l)

(m)

(Note that, from the elastic analog, the initial elastic stress distribution corresponds to the steady

creep solution with n = 1.)

(3) Reference stress analysis

The reference stress is obtained from Eq. (1.10), noting that the limit load in each bar is equal to

Ay

V 1 R

V 2 R

F2A

(n)

(Note that a similar result is obtained by letting n in the stationary creep stress solution as

predicted by the Sim hypothesis.)

The following results are obtained for case #1, where L1 = L2/8:

Initial elastic stress: 1(0) = (8/9)F/A, 2(0) = (1/9)F/A

Stationary creep stress: 1() = (2/3)F/A , 2() = (1/3)F/A

Reference stress: 1(R) = (1/2)F/A, 2(R) = (1/2)F/A

In a similar fashion, for case #2, where L1 = L2/2:

Initial elastic stress: 1(0) = (2/3)F/A , 2(0) = (1/3)F/A

Stationary creep stress: 1() = (0.56)F/A , 2() = (0.44)F/A

Reference stress: 1(R) = (1/2)F/A, 2(R) = (1/2)F/A

Discussion

Case #1 is representative of a highly unbalanced system with an extreme stress concentration. The

initial elastically calculated stresses differ by a factor of 8. The stationary creep stresses differ by a

factor of 2 and the reference stress is equal in both bars. This phenomena is sometimes referred to as

load shedding However, for this highly unbalanced system, the strain in the shorter, stiffer bar is

also a factor of 8 higher than the lower stressed bar. As a result, the question arises as to whether there

ASM_Jawad_Ch01.indd

24

MTC

02/24/2009 02:18AM

Basic Concepts 25

is sufcient creep ductility in the shorter bar to eventually realize the lower reference stress level. If the

shorter bar fails prematurely, the entire load will shift to the longer, more lightly loaded bar, causing

the stress in that bar to increase to twice the reference stress level. Depending on the creep ductility

and how far into the loading cycle the shorter bar fails, the result could be a premature failure of the

two-bar system.

Case #2 represents a more balanced system without an extreme stress concentration. The stationary

creep solution is within 6% of the reference stresses and the strain ratio is only a factor of 2. In this

case, one would not expect a premature failure.

Although it is difcult develop quantied guidance from these two cases, the lesson to be taken from

this is that considerable caution should be taken in applying reference stress methods to highly unbalanced systems, particularly if the creep ductility of the material of construction is suspect.

CYCLIC LOADING

This is a general discussion to acquaint the designer with relevant structural phenomena of signicance to the evaluation of elevated temperature failure modes associated with structural discontinuities and cyclic loading. Specic rules and procedures will be presented in subsequent chapters.

1.6.1

Elastic Follow-Up

Elastic follow-up can cause larger strains in a structure with applied displacement-controlled loading than would be calculated using an elastic analysis. These strain concentrations may result when

structural parts of different exibility are connected in series loaded by an applied displacement and

the exible portions are highly stressed. In order for follow-up to occur, in a two-bar model for example as shown in Fig. 1.15, it is necessary for a lower stressed, exible element to be able to generate

inelastic deformation in a more highly stressed adjacent element. The other requirement is that the

lower stressed remainder of the structure be capable of transmitting a signicant deformation to the

more highly stressed portion of the structure subject to inelastic deformation.

In the two-bar example, a displacement applied to the end of the smaller diameter bar, B, will

initially cause an elastic deformation in both A and B, with B being the more highly stressed bar. Although under creep conditions the stress in A and B will both relax, the higher stress in B will cause

further creep deformation in B and some of the initial elastic deformation in A will be absorbed in B

hence the term elastic follow-up. This process is shown schematically in Fig. 1.16. The path 0-1

is the initial elastic loading and the path 1-2 shows the stress relaxation in time, T. If there were no

elastic follow-up, i.e., the stresses in A and B were equal, there would be pure relaxation without

follow-up along path 1-3.

Also, if A is very much stiffer than B, such that the force in B will cause a negligible deformation in

A, then there will also be pure relaxation in B. This is why the stress due to a local hot spot in a vessel

wall is classied as a peak thermal stress.

One method for dening elastic follow-up is to compute the ratio of 0-2, the creep strain in B at

time, t, to the creep strain that would have occurred under pure relaxation, 0-1. Thus, the elastic follow-up, q, is given by

q

H 0c 2 H 0c 1

(1.11)

Note that for q = 1, there is no follow-up, just pure relaxation and if q the stress in B behaves

as though loaded by a sustained load, i.e., load-controlled, rather than displacement-controlled. For

most geometries loaded in the displacement-controlled mode, q = 2 or less with a reasonable upper

bound of q 3. A more representative case of elastic follow-up is illustrated by Fig. 1.17, a tubesheet

ASM_Jawad_Ch01.indd

25

MTC

02/24/2009 02:18AM

26 Chapter 1

FIG. 1.15

TWO-BAR MODEL WITH ELASTIC FOLLOW-UP

connected to a shell at a different temperature. Elastic follow-up effects can increase the strain and

stress levels at the tubesheet-to-shell junction as shown in Fig. 1.18, which shows representative hysteresis loop with and without elastic follow-up.

The main consequence of elastic follow-up is to reduce the predicted cyclic life as compared to the

life that would be predicted from an elastic analysis without consideration of elastic follow-up. This

is due to two effects as shown in Fig. 1.18. The actual strain range will be greater than that predicted

by an unadjusted elastic analysis and the stress level will be higher due to the slowed rate of stress

relaxation.

In general, there are two approaches to deal with this problem. The rst and most rigorous approach is to do a full inelastic analysis, which predicts the stress and strain at critical points in the

structure as a function of time. The disadvantage of this approach is that it requires complex models

FIG. 1.16

DEFINITION OF ELASTIC FOLLOW-UP

ASM_Jawad_Ch01.indd

26

MTC

02/24/2009 02:18AM

Basic Concepts 27

FIG. 1.17

TUBESHEET-TO-SHELL JUNCTION WITH RELATIVE DEFLECTION, ,

DUE TO TEMPERATURE

of material behavior, which may only have been established for a quite limited number of materials.

These models require substantial judgment in their selection and use, and the actual computation

times and effort involved in interpreting the results can be signicant.

The second approach is based on elastic analysis without directly considering the effects of inelastic

behavior. In this approach, adjustments are made to the elastic analysis results to compensate for the

effects of inelastic behavior. The disadvantage of this approach is that the simpler methods tend to

be overly conservative and the more complex methods can, themselves, be difcult to interpret and

implement.

FIG. 1.18

HYSTERESIS LOOP WITH AND WITHOUT ELASTIC FOLLOW-UP

ASM_Jawad_Ch01.indd

27

MTC

02/24/2009 02:18AM

28 Chapter 1

1.6.2

In Section 1.5.2, the rst step was to separate the structure into free bodies and compute the primary stresses in equilibrium with primary loads. The second step is to establish structural continuity by

applying self-equilibrating loads to the boundaries of the free body segments. The stresses resulting

from these self-equilibrating loads are called discontinuity stresses. This procedure for calculating discontinuity stresses is described in detail in Article A-6000 of the Section III, Division 1 Appendices.

At elevated temperature where creep is signicant, it has been shown by analysis and experiment

that the discontinuity stresses resulting from applied pressure do not relax as might be expected from

self-equilibrating loads. Figure 1.19 (Becht et al., 1989) is a comparison of the analytically predicted

stress history for several structural congurations and loading conditions. A key comparison is case

#5 for a built-in cylinder (radial and rotational constrain at the edge) versus case #1 for pure straincontrolled stress relaxation. After an initial redistribution of stress across the thickness, the discontinuity stress at the built-in edge is essentially constant, analogous to a primary stress. The explanation

for this is that under creep conditions, the free body segments of the structure are undergoing

continuous deformation due to creep with a resultant continuous increasing relative displacement

at the interfaces. This increasing relative displacement prevents relaxation of the interface loads and

resultant discontinuity stresses. Although this phenomenon does not exactly t the elastic follow-up

model, the resulting non-relaxing discontinuity stress is analogous to the case where q , indicative

of a sustained, non-relaxing load, e.g., a primary stress.

Pressure (and mechanical load)-induced discontinuity stresses do not affect the required wall thickness, but they can affect the strain and accumulated creep damage at structural discontinuities. Thus,

in the rules in Subsection NH and most other elevated temperature nuclear code criteria, these stresses

are classied as primary when evaluating strain limits and creep-fatigue damage using the results of

FIG. 1.19

STRESS RELAXATION AND STRAIN ACCUMULATION FOR PRESSURE DISCONTINUITY

(CASE 5), THERMAL DISCONTINUITY (CASE 10), AND SECONDARY STRESS

(CASES 1 AND 2) (RAO, 2002)

ASM_Jawad_Ch01.indd

28

MTC

02/24/2009 02:18AM

Basic Concepts 29

elastic analyses. If strain limits and creep-fatigue damage are evaluated using inelastic analysis accounting for creep, then this effect is automatically included.

For criteria based on Design-by-Rule, there are some restrictions that qualitatively address nonrelaxing pressure-induced discontinuity stresses. For example, in the design rules of Section VIII,

Div 1, this issue is addressed for a number of congurations by the requirement for a 3:1 taper when

joining plates of unequal thickness (UW-9(c)) and head to shell joints (UW-13).

1.6.3

Ratcheting can be described as progressive incremental deformation and shakedown can be described as the absence of ratcheting. Another, similar denition of shakedown is used in the criteria

for Section III, Class 1 and Section VIII, Div 2. In this case, shakedown is considered to occur when

there is negligible plasticity after a few loading cycles. This later approach is illustrated by Figs. 1.20a

and b for elastic-plastic materials with no strain hardening or creep. Consider a tensile specimen that

is strained in tension to a value t, as shown in Fig. 1.20a, which is somewhat greater than the strain

at yield, y, and less than twice y. The initial loading will follow path OAB, initially yielding at A

and continuing to plastically deform until the maximum tensile strain is reached at B. The unloading

portion of the cycle consists of reversing the applied strain to the original staring point, O, following

path BC without yielding. Subsequent loading for the same strain range, 0 t 0, will cycle along

the path BC without yielding, hence the term shakedown. If, on the other hand, the applied strain

range is greater that twice the strain at yield, as shown in Fig. 1.20b, the loading will follow the path

OADEF and there will be yielding from E to F on the unloading cycle. Subsequent cycles of the same

strain range will trace out a hysteresis loop with plasticity at each end of the cycle. What enables

shakedown when the strain range, t, is less than twice the strain at the yield strength is the establishment of a residual stress extending the strain range that can be achieved without yielding in cyclic

strain-controlled loading. However, because the residual stress is limited to the yield strength, if the

applied strain range exceeds twice the strain at yield there will be straining in tension and compression

at either end of the cycle and shakedown will not occur.

In the creep regime, the residual stresses will relax and the strain range that can be achieved without

yielding on each cycle is reduced. There is, however, a quite useful elevated temperature analogy to

FIG. 1.20

STRESS AND STRAIN FOR CYCLIC STRAIN-CONTROLLED LOADING WITHOUT CREEP

ASM_Jawad_Ch01.indd

29

MTC

02/24/2009 02:18AM

30 Chapter 1

the low-temperature shakedown concept. This is illustrated in Figs. 1.21 and 1.23, which are plots of

stress versus time, and Figs. 1.22 and 1.24, which are the corresponding plots of stress versus strain

for strain-controlled loading.

As in the preceding example without creep, in the case shown in Figs. 1.21 and 1.22 the initial

loading follows the path OAB but now the strain is held constant for a certain period and the stress

relaxes to B. The strain is then reversed to point 0, the initial starting point, reaching a compressive

stress at C. If the initial strain beyond yield is sufciently limited, there will be no yielding when the

applied strain is reversed and, assuming that there is no creep at the reversed end of the cycle (either

the temperature is below the creep range or the duration is short), there will not be subsequent yielding upon reloading to point B. During the next tensile hold portion of the cycle, the stress will relax to

B and will subsequently reach C when the strain cycle is reversed. Under these conditions, the cyclic

history will be as shown in Figs. 1.21 and 1.22, and the criteria for shakedown will be that the stress

range associated with the applied strain range does not exceed the yield stress at the cold (or short

duration) end of the cycle plus the stress remaining after full relaxation of the yield stress over the life

of the component. Thus, a criterion for shakedown in the creep regime becomes

H t d V yc E cold V r Ehot

(1.12)

where

Ecold and Ehot = modulus of elasticity at cold end and hot end of the cycle, respectively

t = applied strain range

yc = yield strength at the cold (or short time) end of the cycle, approximately equal to 1.5Scold

r = relaxation strength, the stress remaining after the strain at yield strength is held constant for

the total life under consideration. It can be conservatively approximated by 0.5Shot

Equation (1.12) can also be expressed in terms of calculated stress as

P L Pb Q d V yc V r

(1.13)

where

Pb = primary bending stress intensity

PL = primary local membrane stress intensity

Q = secondary stress

FIG. 1.21

STRESS HISTORY FOR CYCLIC STRAIN CONTROL WITH CREEP AND SHAKEDOWN

ASM_Jawad_Ch01.indd

30

MTC

02/24/2009 02:18AM

Basic Concepts 31

FIG. 1.22

STRESS AND STRAIN CYCLIC STRAIN CONTROL WITH CREEP AND SHAKEDOWN

And PL, Pb, and Q are computed using the values of E corresponding to the operating condition being evaluated. Note that, generally, the use of Ecold will be conservative.

The above is a very useful concept for the design of elevated temperature components, as will be

discussed in more detail in subsequent chapters, particularly Chapters 4 and 5.

Figures 1.23 and 1.24 illustrate the case where the strain range is greater and there is yielding at the

end of the strain unloading cycle. The load path goes from 0 to A with subsequent yielding until D.

During the hold time at D, the stress relaxes to E. This is followed by the unloading or reversed strain

portion of the cycle that results in yielding at F until the cycle is completed at G. For the cycle shown,

FIG. 1.23

STRESS HISTORY FOR CYCLIC STRAIN CONTROL WITH CREEP AND

WITHOUT SHAKEDOWN

ASM_Jawad_Ch01.indd

31

MTC

02/24/2009 02:18AM

32 Chapter 1

FIG. 1.24

CYCLIC STRAIN CONTROL WITH CREEP AND WITHOUT SHAKEDOWN

there will be subsequent yielding on the tensile portion of the cycle and the stress at the start of the

hold period, D, will equal the hot yield strength. In effect, the creep damage for each cyclic hold time

will be reinitiated at the yield strength, thus greatly increasing the creep damage as compared to the

previous case where the stress at the start of the cycle will be the same as the relaxed stress at the end

of the previous hold time. In the case of the higher strain range, a hysteresis loop is established with

creep strain and yielding in each cycle. One of the signicant differences between the two cases is that

for shakedown, the creep damage is only that associated with monotonic stress relaxation throughout

life. In the alternate case, where shakedown is not achieved, the creep damage is accumulated at a

signicantly higher stress level.

As noted above, another use for the term shakedown is to denote freedom from ratcheting, i.e.,

progressive incremental deformation. In the preceding example, the loading was considered to be a

fully reversed strain-controlled cycle. However, in normal design practice, there are both primary loads

and secondary loads and these loadings in combination can cause ratcheting. (Purely displacementcontrolled thermal stresses can also result in ratcheting, but these cases are usually associated with

complete through-the-wall yielding.)

Bree (1968) evaluated potential ratcheting in cylinders under constant pressure and with a cyclic

linear radial thermal gradient. He developed the relationships shown in Fig. 1.25, which identify various regimes of behavior as a function of the relative magnitude of the elastically calculated thermal

stress and pressure stress divided by the yield stress. Six areas of behavior were identied. Regions R1

and R2 resulted in ratcheting or incremental growth even without creep. Region P resulted in plastic

cycling in the absence of creep and the regions S1 and S2 shook down to elastic action after one or two

cycles, again in the absence of creep. If creep is considered, only one region, E, resulted in structural

ASM_Jawad_Ch01.indd

32

MTC

02/24/2009 02:18AM

Basic Concepts 33

FIG. 1.25

BREE DIAGRAM (BREE, 1967)

behavior that could be considered as not being subject to incremental growth. Although useful, the

simple approach of designing to remain in the elastic region, E, can be too restrictive.

Based on a Bree type model, ODonnell and Porowski (1974) developed a less conservative approach to assess the strains accumulated due to pressure (primary stresses) and cyclic thermal gradients (secondary stresses). This technique is a methodology for putting an upper bound on the strains

that can accumulate due to ratcheting. The key feature of this technique is identifying an elastic core

in a component subjected to primary loads and cyclic secondary loads. Once the magnitude of this

elastic core has been established, the deformation of the component can be bounded by noting that

the elastic core stress governs the net deformation of the section. Deformation in the ratcheting, R,

regions of the Bree diagram can also be estimated by considering individual cyclic deformation. The

resulting modications to the basic Bree diagram are shown in Fig. 1.26.

A much more comprehensive development of the Bree and ODonnell/Porowski assessment of

ratcheting is presented in Appendix A.

ASM_Jawad_Ch01.indd

33

MTC

02/24/2009 02:18AM

34

Chapter 1

FIG. 1.26

ODONNELL AND POROWSKIS MODIFIED BREE DIAGRAM (ASME III-NH)

1.6.4

A very important consideration in elevated temperature design is the reduction in cyclic life due to

the effects of creep. This is illustrated by Fig. 1.27 showing the effects of hold time on the cyclic life of

304 stainless steel at 1100F (595C). The loading cycle is constantly increasing tensile strain followed

by a hold time at a xed strain and then a constantly decreasing strain back to the original starting

point. This is referred to as a strain-controlled test as compared to a load-controlled test, where the

load is increased to a xed level and then reversed. During the hold period at a xed strain, the specimen undergoes pure relaxation with no elastic follow-up. As can be seen from the gure, as the hold

ASM_Jawad_Ch01.indd

34

MTC

02/24/2009 02:18AM

Basic Concepts 35

FIG. 1.27

EFFECT OF TENSION HOLD TIME ON THE FATIGUE LIFE OF AISI TYPE 304 STAINLESS

STEEL AT 1100F IN AIR (AMERICAN SOCIETY OF MECHANICAL ENGINEERS, 1976)

time increases the cycles to failure reduce. For a hold time of 1 hour, the reduction in life in this test is

on the order of a factor of 10, and it is not clear from the data if longer hold times will result in a further

reduction in life. Fortunately, for most materials, as the hold time increases the stress relaxes and the

rate damage accumulation slows until the effect essentially saturates. Such a saturation effect is shown

in Fig. 1.28, which is based on 304 stainless steel data at 1200F (650C). At this temperature, relaxation is fairly rapid and saturation occurs in the range of roughly 1 to 10 hours depending on the strain

range. Unfortunately, the actual loads encountered in design are not usually purely strain-controlled

because there are usually follow-up effects and non-relaxing stresses from primary loading conditions.

The development of design methodologies to account for the effect of these varying loading mechanisms is one of the greatest challenges in elevated temperature design.

1.6.4.1 Linear Life Fraction Time Fraction. A number of methods have been explored to correlate creep-fatigue test data and to evaluate cyclic life in design. The method chosen for III-NH is

linear damage summation based on linear life fraction for creep damage and Miners rule for fatigue

damage

'TTr nN f d D

(1.14)

where

DT = time at a given stress level

Tr = allowable time to rupture at that stress level

ASM_Jawad_Ch01.indd

35

MTC

02/24/2009 02:18AM

36

Chapter 1

FIG. 1.28

HOLD-TIME EFFECTS ON FATIGUE LIFE REDUCTION FOR 304 AND 316 STAINLESS

STEEL (AMERICAN SOCIETY OF MECHANICAL ENGINEERS, 1976)

Nf = allowable number of cycles at that strain range

D = factor to account for the interaction of creep and fatigue damage

In practice, safety factors are applied to the given stress level to determine a conservative time to

rupture and to the allowable number of cycles.

For a design evaluation using the above relationship, it is necessary to determine the stress and

strain as a function of time at critical points in the structure. This is conceptually straightforward using

an inelastic analysis that models stress and strain behavior as a function of time. However, as previously noted, the disadvantage of this approach is that it requires complex models of material behavior,

which may only have been established for a quite limited number of materials. These models require

substantial judgment in their selection and use, and the actual computation times and effort involved

in interpreting the results can be signicant.

Alternatively, one can use elastic analysis results in mechanistic models to bound the stress-strain

history without directly considering the effects of inelastic behavior. In this approach, adjustments are

made to the elastic analysis results to compensate for the effects of inelastic behavior. The disadvantage of this approach is that the simpler methods tend to be overly conservative and the more complex

methods can, themselves, be difcult to interpret and implement.

1.6.4.2 Ductility Exhaustion. Another frequently used approach to the assessment of creep damage during cyclic loading is ductility exhaustion. In its simplest form, the combined effect of creep and

fatigue damage may be expressed as

ASM_Jawad_Ch01.indd

36

MTC

02/24/2009 02:18AM

Basic Concepts 37

(1.15)

where

Dc = strain increment

dc = creep ductility

n = number of cycles of a given strain range

Nf = allowable number of cycles at that strain range

Dcf = factor to account for the interaction of creep and fatigue damage

Most of the above comments regarding the application of the time fraction approach also apply

to the ductility exhaustion approach. Conceptually, it should be easier to calculate creep damage

via ductility exhaustion as compared to time fractions because the time to rupture is quite sensitive

to calculated stress; in contrast, creep ductility can be relatively constant depending on the material.

However, in practice there have been numerous modications to the ductility exhaustion approach to

take into account the variations of creep ductility and signicance of when strain accumulation occurs

during the loading cycle. Experimentally, some studies show better correlation for some materials

using ductility exhaustion, depending on selected modications, but there has not been a clear indication of universal applicability. The ductility exhaustion approach tends to see greater use as a damage

assessment tool in failure analyses than as a design tool.

1.7

There are two types of buckling that need to be considered: elastic or elastic-plastic buckling that may

occur instantaneously at any time in life, and creep buckling, which may be caused by enhancement of

initial imperfections with time resulting in geometric instability. The essential difference between elastic and elastic-plastic buckling and creep buckling is that elastic and elastic-plastic buckling occurs with

increasing load independent of time, whereas creep buckling is time-dependent and may occur even

when loads are constant. Elastic and elastic-plastic buckling depends only on the geometric conguration and short-time material response at the time of application. Creep buckling occurs at loads below

the elastic and elastic-plastic buckling loads as a result of creep strain accumulation over time.

The sensitivity of creep buckling to initial imperfections is illustrated by the deformation-time relationships shown in Fig. 1.29. Although typical of the behavior of axially compressed columns and externally pressurized cylinders, these curves are representative of most structures. In general, a structural

FIG. 1.29

CREEP BUCKLING DEFLECTION-TIME CHARACTERISTICS (AMERICAN SOCIETY OF

MECHANICAL ENGINEERS, 1976)

ASM_Jawad_Ch01.indd

37

MTC

02/24/2009 02:18AM

38

Chapter 1

component will deviate initially from a perfect geometrical structure by some small amount. Under

a system of loads, below those that would cause elastic or inelastic instability, the initial deection

is magnied over time due to creep. The deection increases until the geometrical conguration becomes unstable, as shown by point A in Fig. 1.29, and buckling occurs.

In Section VIII, Div 1 and Subsection NH, gures are provided that dene temperature and, in the

case of Subsection NH, time limits within which creep effects need not be considered when evaluating

buckling and instability.

Problems

1. The maximum effective strain in the longitudinal weld of a steam drum is limited to 0.5%. What is

the effective operating stress in the cylinder, from Fig. 1.10, that will result in an expected life of

a. 100,000 hours

b. 300,000 hours

2. What is the allowable stress of the material shown in Figs. 1.11 and 1.12 at 1100F based on

creep and rupture criteria?

3. A pressure vessel component operating at 850F has an expected life of 300,000 hours. What

is the expected life if the temperature was inadvertently raised to 900F while maintaining the

same stress level?

ASM_Jawad_Ch01.indd

38

MTC

02/24/2009 02:18AM

SERVICES, HOUSTON, TEXAS)

ASM_Jawad_Ch02.indd

40

MTC

02/24/2009 04:26AM

CHAPTER

2

AXIALLY LOADED MEMBERS

2.1

INTRODUCTION

Axially loaded members subjected to elevated temperatures are encountered in many structures

such as internal cyclone hangers in process equipment, structural supports for internal trays, and supports of pressure vessels. It is assumed that buckling, which is discussed in Chapter 7, is not a consideration in this chapter. Theoretically, the analysis of uniaxially loaded members operating in the creep

range follow Nortons relationship, correlating stress and strain rate in the creep regime given by

k Vn

dH dT

(2.1)

where

k

= constant

n

= creep exponent, which is a function of material property and temperature

d/dT = strain rate

= stress

This equation, however, is impractical to use for most problems encountered by the engineer. Its

complexity arises from the non-linear relationship between stress and strain rate. In addition, the

equation has to be integrated to obtain strain, and thus deections. A simpler method is normally

used to solve uniaxially loaded members. This method, referred to as the stationary stress method

or the elastic analog method, consists of using a viscoelastic stress-strain equation to evaluate stress

due to creep rather than the more complicated creep equation, which relates strain rate to stress. The

viscoelastic equation is given by

H

K Vn

(2.2)

The results obtained by the stationary stress, Eq. (2.2), are approximate but adequate for most

engineering calculations. This method was rigorously discussed by Hoff (1958) and mentioned in

numerous articles such as those by Hult (1966), Penny and Marriott (1995), and Finnie and Heller

(1959). Hoff proved that Nortons equation (Eq. 2.1) can be replaced with the classical viscoelastic

stress-strain relationship of Eq. (2.2) for a wide range of structures encountered by the engineer when

the following conditions and assumptions are satised:

Primary strain is ignored and only secondary strain is considered.

Strain obtained from Eq. (2.2) is numerically equal to the strain rate in Eq. (2.1).

The material property is in accordance with Eq. (2.1).

The stress eld in the solid must remain constant with time. This is usually achieved after initial stress redistribution in an indeterminate structure due to load or temperature application.

41

ASM_Jawad_Ch02.indd

41

MTC

02/24/2009 04:26AM

42

Chapter 2

Both the viscoelastic structure as well as the creep structure are loaded in the same manner

and have the same boundary conditions.

The justication for using Eq. (2.2) in lieu of Eq. (2.1), with the stated assumptions, can be illustrated for the simple case shown in Fig. 2.1. The stress at sections 1-1 and 2-2 is given by

1 = F/A1

and

2 = F/A2

where

A1, A2 = cross-sectional areas at locations 1-1 and 2-2

1, 2 = stress at locations 1-1 and 2-2

Integration of Eq. (2.1) with respect to time at locations 1-1 and 2-2 and using the above expressions gives

e 1 = k (F/A1) n DT

and

e 2 = k (F/A2) n DT

(2.3)

where

DT = time increment

The ratio of these two expressions is

H 1H 2

FA 1 n FA 2 n.

(2.4)

The same expression given by Eq. (2.4) can also be obtained from Eq. (2.2) at these two locations.

Thus, the relationship between stress and strain in a structure, subject to the assumptions made above,

is the same whether calculated from Eq. (2.1) or Eq. (2.2). Accordingly, Eq. (2.2) is used because it

is more practical to solve. The designer must realize that the strains obtained from Eq. (2.2) actually

correspond to the strain rates obtained from Eq. (2.1).

The application of Eq. (2.2) to regular engineering problems, although simpler than Eq. (2.1), is still

very complicated due to the non-linear relationship between stress and strain. This is illustrated in the

following example.

Example 2.1

What are the forces in the cyclone support rods shown in Fig. 2.2a using (1) elastic analysis,

(2) plastic analysis, and (3) creep analysis? Assume all members to have the same cross-sectional

area, A.

FIG. 2.1

AXIAL TENSION MEMBER

ASM_Jawad_Ch02.indd

42

MTC

02/24/2009 04:26AM

FIG. 2.2

STRUCTURAL FRAME

Solution

(a) Elastic analysis

Referring to Fig. 2.2b, Summation of the forces in the horizontal direction gives

F1

F3

(1)

P 1638 F 1

F2

(2)

'

(3)

F2 40AE

' sin 35

F1 40 cos 35AE

or

'

(4)

85134 F1AE

F2

2128 F1

F2

0565 P

(5)

F1

0266 P

F3

0266 P

ASM_Jawad_Ch02.indd

43

MTC

02/24/2009 04:26AM

Chapter 2

44

TABLE 2.1

F/P VALUES FOR EXAMPLE 2.1

Analysis

Elastic

Creep

F1/P

F2/P

F3/P

1

2

3

4

5

10

100

0.266

0.266

0.323

0.342

0.351

0.357

0.368

0.378

0.379

0.565

0.565

0.471

0.440

0.424

0.415

0.397

0.381

0.379

0.266

0.266

0.323

0.342

0.351

0.357

0.368

0.378

0.379

Plastic

In a perfectly plastic analysis, it is assumed that all members have reached their yield stress value.

Background of the structural plastic theory may be found in various books such as Baker and Heyman

(1969) and Beedle (1958). Summation of the forces (Fig. 2.2d) in the vertical direction gives

2 sA A (cos 35 ) + sA A = P

or

F1

F2

F3

0.379 P

A comparison between the maximum force in parts (a) and (b) shows an almost 50% reduction. The

results are shown in the last row of Table 2.1.

(c) Creep analysis

Equations (1) and (2) using summation of forces are still applicable

F1

F3

V1

V3

or

(6)

and

F2

P 1638 F1

V2

PA 1638 V 1

or

(7)

Equations (3) and (4) cannot be used in creep analysis because the relationship D = FL/AE applies

only in the elastic range. The strain in members (1) and (2) are expressed as

e 1 = D (sin 35) /L1 = D(sin 35) /48.831

(8)

and

H2

'400

(9)

The stationary stress Eq. (2.2) will have to be used to correlate stress and strain. Substituting Eqs.

(7), (8), and (9) into Eq. (2.2) results in the following two expressions

' sin 3548831

K V1n

(10)

and

'40

ASM_Jawad_Ch02.indd

44

K PA 1638 V 1 n

MTC

(11)

02/24/2009 04:26AM

s1 [1.638 + (2.128)1/n ] = P/A

F 1 = [1.638 + (2.128 ) 1/n ]-1 P

F 2 = {1 - 1.638 [1.638 + (2.128 ) 1/n ] -1} P

F 3 = [1.638 + ( 2.128 ) 1/n ]-1 P

These expressions are substantially more complex than those obtained from the elastic or plastic

analysis. The complexity increases exponentially as the number of members and degrees of freedom

increase to the point where it is impractical to use this manual approach to solve large problems.

The magnitudes of F1 and F2 for various n values are shown in Table 2.1. For n = 1, the values of

F1 and F2 are the same as those obtained from an elastic analysis. Also, as n approaches innity, the

values of F1 and F2 approach those obtained from a plastic analysis. Most materials used for pressure

vessel construction operating at elevated temperatures have an n value in the neighborhood of 2.5 to

10. Table 2.1 shows that the difference between the maximum member force having n = 5 and that

obtained from plastic design is only about 9%. This difference reduces to about 5% for n = 10. Accordingly, many engineers in the boiler and pressure vessel area use plastic analysis to evaluate structures in the creep range taking into consideration the accuracy of the analysis. Also, plastic analysis is

much simpler to use compared to creep analysis. The designer can use Eq. (2.2) or a non-linear nite

element analysis when a more accurate result is needed.

Table 2.1 also shows that high member forces obtained from an elastic analysis tend to reduce in

magnitude due to creep, whereas low member forces tend to increase in magnitude due to creep. Accordingly, designing members with high forces obtained from an elastic analysis at elevated temperature results in a conservative design. Conversely, designing members with low forces obtained from an

elastic analysis at elevated temperature results in an unconservative design.

It must also be kept in mind that both the creep and plastic analyses result in some member forces

that change from tension to compression as the applied loads are reduced or eliminated during a cycle.

Accordingly, a special consideration must be given to bracing.

AND VIII-1 AS A GUIDE

Design of axially loaded structural components at elevated temperatures in ASME I and VIII-1

construction is straightforward. These two codes require an elastic analysis with the allowable stress

obtained from II-D. At elevated temperatures, the allowable stress in II-D is based on the creep and

rupture criteria as discussed in Chapter 1. Hence, at elevated temperatures, ASME I and VIII-1 permit

an elastic analysis with the allowable stresses obtained from creep and rupture data. This approach,

although commonly used, has many drawbacks. These drawbacks include the lack of imposing limits

on strain and deformation due to creep and ignoring thermal stresses that may lead to excessive strain

in the creep regime due to cyclic conditions.

As described in Section 1.6.3, there is a quite useful concept in the creep regime that is analogous

to the shakedown concept below the creep regime. The resultant shakedown concept in the creep

regime is given by Eq. (1.12) as

H t d V yc E cold V r Ehot

(2.5)

where

Ecold and Ehot = modulus of elasticity at cold end and hot end of the cycle, respectively

t = applied strain range

ASM_Jawad_Ch02.indd

45

MTC

02/24/2009 04:26AM

46

Chapter 2

yc = yield strength at the cold (or short time) end of the cycle; it is approximately equal to 1.5Scold

r = relaxation strength, the stress remaining after the strain at yield strength is held constant for

the total life under consideration; it can be conservatively approximated by 0.5Shot

Equation (2.5) can also be expressed in terms of calculated stress as

P L Pb Q d V yc V r

(2.6)

where

Pb = primary bending equivalent stress

PL = primary local membrane equivalent stress

Q = secondary equivalent stress

And PL, Pb, and Q are computed using the values of E corresponding to the operating condition being

evaluated. Note that, in general, the use of Ecold will be conservative. For axial members, Pb is zero

and Eq. (2.6) becomes

PL Q d 15 S C S H 2

(2.7)

It is of interest to note that the axial equation for stresses below the creep range is given in VIII-2

as

PL Q d 15S C 15S H

(2.8)

Equation (2.7) was developed by ASME with the intent of assuring shakedown after a few cycles.

A comparison of the creep (2.7) and the VIII-2 Eq. (2.8) show them to be similar with the exception

of the last term.

The analysis of trusses is easily done by computers using matrix structural analysis. A brief introduction to the application of the matrix theory in solving axial member problems is presented in this

section.

Figure 2.3 shows a typical truss member. Positive direction of forces, applied loads, joint deections, and angles are shown in Fig. 2.3. Member 1, shown in the gure, starts at node 1 and ends at

node 2. At node 1, forces P1 and P2, or deections X1 and X2 in the x and y directions, may be applied

as shown in the gure. Similarly, forces P3 and P4, or deections X3 and X4, can also be applied at node

2. The local member rotation angle with respect to global coordinate system of the truss is measured

FIG. 2.3

SIGN CONVENTION FOR AXIAL MEMBERS

ASM_Jawad_Ch02.indd

46

MTC

02/24/2009 04:26AM

positive counterclockwise from the positive x direction at the beginning of the member. Tensile force

is as shown in the gure.

The general relationship between applied joint loads and forces in a truss is given by

>P @

>K @>X @

(2.9)

where

[P ] = applied loads at truss joints

[K ] = global stiffness matrix of truss

[X ] = deformation of truss joints

The global stiffness matrix of the truss (Weaver and Gere, 1990) is assembled from the individual

truss member stiffnesses [k ]. The stiffness matrix, [k ], of any member in a truss is expressed as

k1

>k @

k2

k2

k3

k 1 k 2

k 2 k 3

k 1

k 2

k1

k2

k 2

k 3

k2

k3

(2.10)

where

k 1 = AE cos2/L

k 2 = AE sin cos/L

k 3 = AE sin2/L

Equation (2.10) is a 4 4 matrix because there are a total of 4 degrees of freedom at the ends of the

member. They are X1, X2, X3, and X4, or P1, P2, P3, and P4. The stiffness matrix of the truss is assembled

by combining the stiffness matrices of all the individual members into one matrix, called the global

matrix. Once the global stiffness matrix [K ] is developed, Eq. (2.9) can be solved for the unknown joint

deections [X ]. This assumes that the load matrix [P] at the nodal points is known.

>X @

(2.11)

The forces in the individual members of the truss are then calculated from the equation

>F @

>D@> Xm@

(2.12)

where

[F ] = force in member

[D] = force-deection matrix of a member

[Xm] = matrix for deection at ends of individual members

The matrix [D] is dened as

>D @

> D 1 D 2

D1

D 2@

(2.13)

where

D1 = AE cos/L

D2 = AE sin/L

The following example illustrates the solution of a truss problem operating in the creep range using

elastic analysis as an approximation.

ASM_Jawad_Ch02.indd

47

MTC

02/24/2009 04:26AM

48

Chapter 2

Example 2.2

An internal cyclone is supported by three braces as shown in Fig. 2.4a. All braces have the same

cross-sectional area. The total weight of the cyclone and its contents is 10,000 lb. Determine the forces

in the braces and their size. The material of the braces is 304 stainless steel. The design temperature

is 1450F and the modulus of elasticity is 18,700 ksi. The allowable stress in accordance with VIII-1

is 1800 psi at 1450F and 20,000 psi at room temperature.

Solution

Because all members have the same cross-section, they will be assumed to have a cross-sectional

area of 1.0 in.2. The actual area will be determined when the member forces in the truss are obtained

from the analysis. To simplify the discussion in this chapter, and to compare the results with subsequent examples, the braces are assumed to be fabricated from round rods. In actual practice, pipes,

channels, or angles are used.

The load on the truss and the angle of orientation of each truss member are shown in Fig. 2.4b. The

global stiffness is obtained by combining all local member stiffnesses.

Member I (nodal points 1, 2, 3, and 4)

= 60, L = 69.28 in. From Eq. (2.10), the stiffness matrix for member 1 is

116 879

202439

sym m e tri c

67480

116879

67480

116879

202439

116879

2 0 2439

>k @I

67480

FIG. 2.4

CYCLONE SUPPORT FRAME

ASM_Jawad_Ch02.indd

48

MTC

02/24/2009 04:26AM

= 90, L = 60.00 in. From Eq. (2.10), the stiffness matrix for member 2 is

0

-311,667

0

311,667

0

0

0

symmetric

0

311,667

[k ]II =

= 120, L = 69.28 in. From Eq. (2.10), the stiffness matrix for member 3 is

67,480

116 879

67480

sy m m e tri c

116 879

202439

116879

2 0 2 439

>k @III

116,879

202,439

67480

Combining the above three matrices and deleting the terms that pertain to nodal points 3 through 8

in Fig. 2.4b because they are attached to the head, and hence xed against movement, gives

134960

716545

>P @

>K @

0

10000

798 u 10 8

140 u 10 2

X2

X1

3262 lb

F3

3262 lb

4350 lb

F1

The maximum force in the support system is in member 2 and is equal to 4350 lb. The required area

is

A

4350 1800

242 in.2

ASM_Jawad_Ch02.indd

49

MTC

02/24/2009 04:26AM

50

Chapter 2

d 2/4 = 2.42

d = 1.75 in.

or

2

A = 2.42 in. .

The elastic stresses in members 1 through 3 are then

1350

1350

1800

psi

The effect of temperature variation on the member forces in a truss can easily be incorporated into

the matrix equations as well. The procedure consists of calculating the following equivalent joint load

for members that have a change in temperature

(2.14)

D 'T AE

where

A

= cross-sectional area of member

E

= modulus of elasticity

F

= equivalent member force

(DT ) = change of temperature in member

The equivalent F force is then entered as a nodal force and the truss nodal deections are calculated

from Eq. (2.11). The nal forces are those calculated from Eq. (2.12) minus the quantity from Eq.

(2.14) for those members that have temperature changes.

For axial members, it is assumed that the mechanical stresses are primary in nature and thermal

stresses are secondary in accordance with the general criteria of VIII-2. The analysis procedure is illustrated in the following example.

Example 2.3

Member 2 in Fig. 2.4 is subjected to a 25F increase in temperature excursion. What are the forces

in all members due to this temperature increase? Let = 10.8 10-6 in./in.F, E = 18,700 ksi. Assume

the area for all members to be A = 2.42 in.2. The yield stress is 30,000 psi at room temperature and

10,600 psi at 1450F. The allowable stress is 20,000 psi at room temperature and 1800 psi at 1450F.

Solution

The global stiffness is obtained by combining all local member stiffnesses to give

326,600

0

0

1,734 ,000

[K ] =

F

12220 lb

ASM_Jawad_Ch02.indd

50

0

12220

>P @

MTC

02/24/2009 04:26AM

0

0705 u 10

X

X

2

3987

6905

3987

F1

F

F

lb

And the corresponding thermal stress, based on a cross-sectional area of 2.42 in.2, is

1650

1650

2850

psi

It is of interest to note that a small temperature increase in one of the members results in member

stresses that are approximately of the same magnitude as those obtained from a large applied mechanical load. This condition illustrates the importance of maintaining a constant temperature, whenever

possible, in all members during operation.

Total sum of mechanical stresses, PL, from Example 2.2 and thermal stress, Q, from above is

3000

P L Q

3000

1050

psi

The ASME VIII-2 criteria for acceptable stress levels require that

1. For temperatures below the creep range, the mechanical plus thermal stress are less than 3

times the allowable stress or 2 times the average yield stress, whichever is greater. The allowable and yield stress are taken as the average value at the high and low temperature extremes of

the cycle. This criterion, given by Eq. (2.8), is to assure stress shakedown.

2. For temperatures above the creep range, the mechanical plus thermal stress are less than 3

times the allowable stress. The allowable stress is taken as the average value at the high and

low temperature extremes of the cycle. This criterion, given by Eq. (2.7), is to assure stress

creep.

In this example the temperature at one extreme of the cycle is in the creep range and Eq. (2.7) is

applicable. The right-hand side of this equation is

1520000 18002

30900 psi

All (PL + Q) values are well below the allowable stress of 30,900 psi.

AS A GUIDE CREEP LIFE AND DEFORMATION LIMITS

Axially loaded members in ASME I and VIII-1 are occasionally required to be analyzed for fatigue

and creep at elevated temperatures. The analysis is usually based on commonly accepted practice

ASM_Jawad_Ch02.indd

51

MTC

02/24/2009 04:26AM

52

Chapter 2

because there are no published standards that cover this area. The procedure for assessing creep rupture life of a structural member is,

1. Calculate the strain corresponding to the maximum stress in the member due to mechanical

and thermal stresses in accordance with the equation

'H

2Salt E

(2.15)

where

E = modulus of elasticity at the maximum metal temperature experienced during the cycle

2Salt = maximum stress range during the cycle due to mechanical and thermal loads

D = maximum equivalent strain range

2.

n Nf j

(2.16)

where

(n)j = number of applied repetitions of cycle type, j

(Nf)j = number of design allowable cycles for cycle type, j, obtained from a design fatigue table such

as that shown in Table 2.2

3. Calculate the mechanical stress PL.

4. Enter an isochronous chart similar to the one shown in Fig. 2.5 with stress PL and determine

the strain corresponding to the number of expected life hours. The obtained strain should not

exceed 1%.

TABLE 2.2

DESIGN FATIGUE STRAIN RANGE, t, FOR 304 STAINLESS STEEL (ASME, III-NH)

Number of

cycles, Nd

[Note (1)]

10

20

40

U.S. customary units

100F

800F

900F

1000F

1100F

1200F

1300F

0.051

0.036

0.0263

0.050

0.0345

0.0246

0.0465

0.0315

0.0222

0.0425

0.0284

0.0197

0.0382

0.025

0.017

0.0335

0.0217

0.0146

0.0297

0.0186

0.0123

102

2 102

4 102

0.018

0.0142

0.0113

0.0164

0.0125

0.00965

0.0146

0.011

0.00845

0.0128

0.0096

0.00735

0.011

0.0082

0.0063

0.0093

0.0069

0.00525

0.0077

0.0057

0.00443

103

2 103

4 103

0.00845

0.0067

0.00545

0.00725

0.0059

0.00485

0.0063

0.0051

0.0042

0.0055

0.0045

0.00373

0.0047

0.0038

0.0032

0.00385

0.00315

0.00263

0.00333

0.00276

0.0023

104

2 104

4 104

0.0043

0.0037

0.0032

0.00385

0.0033

0.00287

0.00335

0.0029

0.00254

0.00298

0.00256

0.00224

0.0026

0.00226

0.00197

0.00215

0.00187

0.00162

0.00185

0.00158

0.00138

105

2 105

4 105

0.00272

0.0024

0.00215

0.00242

0.00215

0.00192

0.00213

0.0019

0.0017

0.00188

0.00167

0.0015

0.00164

0.00145

0.0013

0.00140

0.00123

0.0011

0.00117

0.00105

0.00094

0.0019

0.00169

0.00149

0.0013

0.00112

0.00098

0.00084

106

NOTE:

(1) Cyclic strain rate: 1 103 in./in./sec. (1 103 m/m/s).

ASM_Jawad_Ch02.indd

52

MTC

02/24/2009 04:26AM

5. Obtain the number of hours, Tm, corresponding to mechanical stress, PL, from a stress-torupture table similar to the one shown in Table 2.3.

6. Calculate the rupture ratio

Tr Tm k

(2.17)

where

Tr = number of required time for the part

7. The fatigue ratio calculated in step 2 and the rupture ratio calculated in step 6 are combined in

accordance with the following creep-fatigue equation

(2.18)

FIG. 2.5

AVERAGE ISOCHRONOUS STRESS-STRAIN CURVES (ASME, III-NH)

ASM_Jawad_Ch02.indd

53

MTC

02/24/2009 04:26AM

54

Chapter 2

TABLE 2.3

EXPECTED MINIMUM STRESS-TO-RUPTURE VALUES, 1000 PSI, TYPE 304

STAINLESS STEEL (ASME, III-NH)

U.S. customary units

Temp.,

F

1 hr

800

850

900

950

1000

1050

1100

1150

1200

1250

1300

1350

1400

1450

1500

57

56.5

55.5

54.2

52.5

50

45

38

32

27

23

19.5

16.5

14.0

12.0

3 102

3 103

3 104

3 105

3

4

5

10 hr 30 hr 10 hr

hr

10 hr

hr

10 hr

hr

10 hr

hr

2

57

56.5

55.5

54.2

50

41.9

35.2

29.5

24.7

20.7

17.4

14.6

12.1

10.2

8.6

57

56.5

55.5

51

44.5

37

31

26

21.5

17.9

15

12.6

10.3

8.8

7.2

57

56.5

55.5

48.1

39.8

32.9

27.2

22.5

18.6

15.4

12.7

10.6

8.8

7.3

6.0

57

56.5

51.5

43

35

28.9

23.9

19.3

15.9

13

10.5

8.8

7.2

5.8

4.9

57

56.5

46.9

38.0

30.9

25.0

20.3

16.5

13.4

10.8

8.8

7.2

5.8

4.6

3.8

57

50.2

41.2

33.5

26.5

21.6

17.3

13.9

11.1

8.9

7.2

5.8

4.7

3.8

3.0

57

45.4

36.1

28.8

22.9

18.2

14.5

11.6

9.2

7.3

5.8

4.6

3.7

2.9

2.4

51

40

31.5

24.9

19.7

15.5

12.3

9.6

7.6

6.0

4.8

3.8

3.0

2.3

1.8

44.3

34.7

27.2

21.2

16.6

13.0

10.2

8.0

6.2

4.9

3.8

3.0

2.3

1.8

1.4

39

30.5

24

18.3

14.9

11.0

8.6

6.6

5.0

4.0

3.1

2.4

1.9

1.4

1.1

where

Dcf = total creep fatigue damage factor obtained from Fig. 2.6.

It is assumed in the above analysis that thermal stresses diminish relatively quickly when calculating

rupture life Tm. It is also assumed that stress concentration factors due to holes, lugs, etc., are included

in the strain values when entering Table 2.2, but are excluded when entering the stress values in Table

2.3. Stress concentration factors may be obtained from various publications such as Pilkey (1997) or

by experimental or theoretical methods.

Example 2.4

The support system in Fig. 2.4 is subjected to the cycles shown in Fig. 2.7. The system temperature

increases from ambient to 1450F. It stays at 1450F for about 3 years (30,000 hours). The temperature is then dropped down to ambient temperature where the system is inspected before it starts up

again. The process is repeated ten times during the expected life of the system (300,000 hours). Evaluate the creep fatigue property of the support system based on the mechanical stress obtained from

Example 2.2 plus the thermal stress obtained from Example 2.3. Let E = 18,700 ksi at 1450F.

Solution

1. Calculate the strain corresponding to the maximum stress in the member due to mechanical

and thermal stresses in accordance with Eq. (2.15).

Summary of the stresses from Examples 2.2 and 2.3 are shown in Table 2.4.

From Eq. (2.15), the maximum strain is

'H

_ 3000 _ 18700000

Total number of cycles n

ASM_Jawad_Ch02.indd

54

10

MTC

02/24/2009 04:26AM

FIG. 2.6

CREEP-FATIGUE DAMAGE ENVELOPE (ASME, III-NH)

Table 2.2 is limited to 1300F. At this temperature, a life of 1,000,000 cycles is obtained for

a strain level of 0.00084. The strain level calculated above is 0.00016, which is about 5 times

smaller. Accordingly, we can use the stain value of 0.00084 in./in. as a starting number for

1,000,000 cycles at 1300F. This strain value is much larger than the calculated value of 0.00016

and is thus very conservative. From Eq. (1.3), the Larson-Miller parameter is

PLM

45,760

FIG. 2.7

OPERATING CYCLES

ASM_Jawad_Ch02.indd

55

MTC

02/24/2009 04:26AM

56

Chapter 2

TABLE 2.4

STRESS SUMMARY (PSI)

Member

from Example 2.2

1

2

3

Example 2.3

1350

1800

1350

1650

-2850

1650

stress

3000

-1050

3000

45760

log10 N f

396 and N f

9080 cycles

nNf

109080 | 0

From Table 2.4, maximum PL = 1800 psi.

4. Check isochronous curve

From Fig. 2.5 with PL = 1800, a value of = 0.4% is obtained. This value is well below the

limiting value of 1%.

5. Obtain the number of hours, Tm, from rupture curve.

From Table 2.3, the 1450F line gives a life of 100,000 hours for a stress of 1800 psi.

6. Calculate the rupture ratio from Eq. (2.17).

Tr Tm

Because this ratio is substantially greater than 1.0, the test fails and the diameter of 1.75 in.

(area = 2.42 in.2) for the rods is not adequate for 300,000 hours.

Second trial

From Table 2.3, the maximum stress PL cannot exceed 1400 psi at 1450F in order to obtain 300,000

hours. Thus, the area of the rods must be increased to

A

The new stresses are shown in Table 2.5.

TABLE 2.5

STRESS SUMMARY (PSI)

Member

from Example 2.2

1

2

3

ASM_Jawad_Ch02.indd

1040

1385

1040

56

Thermal stress, Q,

from Example 2.3

1650

-2850

1650

MTC

stress

2690

-1465

2690

02/24/2009 04:26AM

5. Obtain the number of hours, Tm, from rupture curve.

From Table 2.3, the 1450F line gives a life of 300,000 hours at a stress of 1400 psi.

Tm

300000 hours

Tr Tm

From Fig. 2.5 with a value of (n/Nd) = 0.0 and a value of (Tr/Tm) = 1.0, it can be shown that

D is within the envelope for 304 stainless steel. Hence, the structural members are adequate for

10 cycles and an expected life of 300,000 hours.

Use 2.0-in. diameter rods.

2.4

Many structural components in a boiler and pressure vessel operating in the creep range can be

analyzed, as a limiting case, by plastic analysis as demonstrated by the results in Table 2.1. The reference stress method, discussed in Section 1.5.3.2, is based on the following equation to determine the

equivalent creep stress in a structure

VR

PPL V y

(2.19)

where

P = applied load on the structure

PL = the limit value of applied load based on plastic analysis

R = reference stress

y = yield stress of the material

Plastic analysis results in a larger load-carrying capacity of the indeterminate structure than an elastic analysis. In plastic analysis (Wang, 1970), the structure is loaded until the member with the highest stress reaches a specied limiting stress. Additional increment in the applied loading is assumed

possible subsequent to removing the member with the limiting stress from further consideration. The

analysis continues through various cycles. At the end of each cycle, members that have reached their

limiting stress are removed from further analysis and the process continues by increasing the load

until the structure becomes unstable. The structure is assumed to support the entire load accumulated

through the various cycles. In many instances this load is substantially larger than that used in elastic

analysis and is the limiting case of creep analysis.

Example 2.5

In Example 2.2, let the cross-sectional area of all members = 2.42 in.2. The total weight of the

cyclone and its contents is 10,000 lb. Determine the forces in the braces using the reference stress

method. The material of the braces is 304 stainless steel. The design temperature is 1450F and the

modulus of elasticity is 18,700 ksi. Assume an allowable stress of 1800 psi and average yield stress of

20,300 psi.

ASM_Jawad_Ch02.indd

57

MTC

02/24/2009 04:26AM

58

Chapter 2

Solution

The maximum load capacity per brace using yield stress as a criterion is

Load brace

20300 242

49126 lb

The limit value of the applied load, Plim, based on plastic analysis is calculated as

Plim

134200 lb

The stress in all of the truss members in the creep range is then calculated from Eq. (2.19) as

V R 1000013420020300 1510 psi

This stress value is smaller than the maximum allowable stress value of 1800 psi obtained elastically

for member 2 in Example 2.2. Thus, the designer has two choices at this time. One choice is to use the

new lower stress of 1510 psi in calculating creep-fatigue cycles. The other choice it to reduce the size

of the members based on the lower stress. In this case, the required brace area is (1510/1800)(2.42)

= 2.03 in.2.

Use 1.625-in. rods.

A = 2.07 in. .

Thus, in this case, the reference stress method results in a 17% reduction of required area (2.42/2.07)

compared to elastic analysis using the same factor of safety.

Problems

2.1 An internal cyclone is supported by three braces as shown. All braces have the same crosssectional area of 2.0 in.2. The total weight of the cyclone and its contents is 15,000 lb. The material of the braces is 304 stainless steel. The design temperature is 1450F and the modulus

of elasticity is 18,700 ksi. The allowable stress in accordance with VIII-1 is 1800 psi at 1450F

and 20,000 psi at room temperature. Member 3 is subjected to a 25F increase due to insulation problems. Let = 10.8 10-6 in./in.F. The yield stress is 30,000 psi at room temperature and 10,600 psi at 1450F. Check the following:

a) Stresses due to mechanical loads

b) Stresses due to mechanical and thermal loads in accordance with VIII-2 criteria

2.2 Use stresses obtained from problem 2.1 to evaluate the creep-rupture property for 30,000

hours and one cycle. Use the properties given in Example 2.4.

ASM_Jawad_Ch02.indd

58

MTC

02/24/2009 04:26AM

ASM_Jawad_Ch03.indd

60

MTC

02/24/2009 04:59AM

CHAPTER

3

MEMBERS IN BENDING

3.1

INTRODUCTION

Many components in pressurized equipment operating in the creep range consist of beams and at

plates in bending. These include piping components, internal catalyst supports, nozzle covers, and

tubesheets. In this chapter, the characteristics of beams in bending are studied rst, followed by an

evaluation of at circular plates in bending.

The characteristics of beams in bending operating in the creep range have been studied by many

engineers and researchers. Theoretically, the analysis is based on the following Nortons relationship

correlating stress and strain rate in the creep regime as discussed in Chapters 1 and 2 and as given by

k Vn

dH dT

(3.1)

where

k

n

d/dT

= constant

= creep exponent, which is a function of material property and temperature

= strain rate

= stress

This equation, however, is impractical to use for most problems encountered by the engineer as

discussed in Section 2.1. A simpler method, referred to as the stationary stress or the elastic analog

method, is normally used to solve beam bending problems and is given by

H

K Vn

(3.2)

where

K = constant

= strain

It should be noted that when n = 1.0 in Eq. (3.2), then K =1/E, where E is the modulus of elasticity. The application of Eq. (3.2) to regular engineering problems, although simpler than Eq. (3.1), is

still very complicated due to the non-linear relationship between stress and strain. This difculty is

overcome by performing plastic analysis to approximate the results obtained by Eq. (3.2) as explained

later in the chapter.

The equations for the bending of a beam in the creep regime are based on assumptions (Finnie and

Heller, 1959) that are similar to those for beams in elastic analysis, i.e.,

61

ASM_Jawad_Ch03.indd

61

MTC

02/24/2009 04:59AM

62

Chapter 3

Plane cross-sections before bending remain plane after bending.

Bending deection is small compared to the length of the beam.

Stress-strain diagrams are the same for both the tensile as well as the compressive sides

of the beam.

The plane of bending is a plane of symmetry.

The strain at a given point in a beam due to bending is

H

z U

(3.3)

where

z = location of point measured from the neutral axis

= radius of curvature

The beam is assumed to have achieved a stationary stress condition. This assumption is realistic for

components in pressure vessel applications where temperature and loading are constant for extended

periods of operation during the cycle. Combining Eqs. (3.2) and (3.3) results in

V

1n

z U 1n

(3.4)

Equating the internal moment in a cross-section to the applied external moment gives

V zdA

(3.5)

where

dA = unit cross-sectional area of the beam

M = applied external moment

Combining Eqs. (3.4) and (3.5) yields

M

1 n 1n

U

z1 1n

dA

(3.6)

In

z1 1n d A

(3.7)

Mz 1nIn

(3.8)

where

In = creep moment of inertia

And combining Eqs. (3.4) and (3.6) gives

V

3.2.1

Rectangular Cross-Sections

For a rectangular cross-section with B = width and H = height, Eq. (3.7) gives

H 2

In

z 1 1n dz

H 2

ASM_Jawad_Ch03.indd

62

MTC

02/24/2009 04:59AM

Members in Bending 63

or

2nB

H22 H21n

1 2n

In

(3.9)

Substituting this expression into Eq. (3.8) and dening the elastic moment of inertia as

BH 312

results in

MH

2I

s =

1 + 2n

3n

(3.10)

1/n

z

H/ 2

(3.11)

6MBH 2

for n 10

(3.12)

The rst bracketed term in Eq. (3.11) is the conventional (Mc/I) expression in elastic analysis.

The second and third bracketed terms are the stress modiers due to creep. When n = 1, Eq. (3.11)

gives the conventional elastic bending equation. Figure 3.1 shows the stress distribution in the beam

cross-section when n = 1, 3, 6, and 10. Notice the reduction in maximum stress due to the creep function n.

3.2.2

Circular Cross-Sections

For tubes having circular cross-sections with Ro = outside radius and Ri = inside radius, Eq. (3.7)

gives (Finnie and Heller, 1959),

In

8I

1 Ri Ro 3 1n

* >1 12n@

>3 1n@ S 1 2

1 Ri Ro 4

* >15 12n@

(3.13)

where

I = moment of inertia of a tubular cross-section

= (Ro4 - Ri4)/4

G = gamma function tabulated in Table 3.1

The bending equation becomes

V

(3.14)

For thin shells and pipes, this equation reduces to

MS tR2o

for n

10

(3.15)

Example 3.1

A beam in a high temperature application is subjected to a 1000 ft-lb maximum bending moment.

The allowable stress is 2000 psi. Determine

ASM_Jawad_Ch03.indd

63

MTC

02/24/2009 04:59AM

64

Chapter 3

FIG. 3.1

STRESS DISTRIBUTION IN A RECTANGULAR BEAM AS A FUNCTION OF n

(a) The required cross-section of a rectangular beam if an elastic analysis was performed where n =

1.0.

(b) The required cross-section of a rectangular beam if a creep analysis was performed where n =

4.7.

(c) The required cross-section of a pipe beam if an elastic analysis was performed where n = 1.0.

(d) The required cross-section of a pipe beam if a creep analysis was performed where n = 4.7.

Solution

(a) From Eq. (3.12)

BH 2

61000 u 122000

36 in 3

(b) From Eq. (3.11) with z = H/2,

ASM_Jawad_Ch03.indd

64

MTC

02/24/2009 04:59AM

Members in Bending 65

TABLE 3.1

SOME TABULATED VALUES OF (n)

(n)

(n)

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.0

1.0000

0.9514

0.9182

0.8975

0.8873

0.8862

0.8935

0.9086

0.9314

0.9618

1.0000

NOTE:

x n1

(1) ( n)

e x

dx

2000

1000 u 12H

2I

1 247

347

BH 2 = 48.81 in.3.

(c) From Eq. (3.15),

2

tRo

191 in3

Use 8 in. Sch 5S pipe (OD = 8.625 in., t = 0.109 in., and tRo2 = 2.03 in3.).

FIG. 3.2

THICK CYLINDRICAL SHELL

ASM_Jawad_Ch03.indd

65

MTC

02/24/2009 04:59AM

Chapter 3

66

(d) An initial assumption of using 8 in. Sch 5S pipe resulted in a stress of more than 5000 psi, which

is well above the allowable stress of 2000 psi.

Assume an 8-in. Sch 10S pipe.

OD

Ro

t

Ri

8625 in

43125 in

S 43125 4 41645 4 4

Ri Ro

09657 1n

0148 in

41645 in

3541 in 4

02128 12n

01064

In

83541

1 09657 32128

*>11064 @

32128 S 1 2

1 09657 4

*>16064 @

4293 in 4

V

1205 psi 2000 psi

3.3

SHAPE FACTORS

The shape factor is dened as the ratio of the moment that a cross-section can withstand when

analyzed using creep or plastic analysis to the moment that the cross-section can withstand using

elastic analysis. In this section, the shape factors for rectangular and hollow circular cross-sections

are derived based on creep and plastic analysis. These shape factors will be used later in Chapters 4

and 5 when analyzing shells.

3.3.1

Rectangular Cross-Sections

For rectangular cross-sections, the relationship between maximum stress and moment in the elastic

range is

V

6MBH 2

(3.16)

where

B = width of beam

H = height of beam

M = applied bending moment

= maximum bending stress

The relationship between maximum stress and moment in the creep range is obtained from Eq.

(3.11). Substituting z = H/2 and I = BH 3/12 results in

ASM_Jawad_Ch03.indd

66

MTC

02/24/2009 04:59AM

Members in Bending 67

6M

BH 2

1 2n

3n

(3.17)

6M

BH 2

(3.18)

or

1

SF

where

SF = shape factor

3n

1 2n

(3.19)

For n = 1, the shape factor is 1.0, and Eqs. (3.16) and (3.18) become the same. In the creep range,

n varies for different materials and temperatures. Values of n are published for a number of ferrous

and non-ferrous alloys at different temperatures (Odqvist, 1966). The n values range between 2.5 for

annealed carbon steel at 1020F and 10 for aluminum at 300F. Hult (1966) published similar data.

Variation in SF with respect to n is shown in Table 3.2.

The ASME uses an SF of 1.25 for calculations made in the creep range. This is based in a conservative value of n = 2.5.

Table 3.2 shows that the value of the shape factor is equal to 1.5 as n approaches innity. It so happens that the 1.5 shape factor also corresponds to that obtained from plastic analysis of beams. This is

illustrated in the stress-strain diagram of Fig. 3.3a for an elastic-perfectly plastic material. The stress

and strain distribution increases proportionately with an increase in the external moment as shown

in Fig. 3.3b, points 1 through 4. The stress distribution across the thickness gradually changes from

triangular to rectangular in shape as the strain increases from points 1 to 2, and nally to point 4. In

the fully plastic region at point 3, the stress expression is obtained by equating internal and external

moments

V

4MBH 2

(3.20)

A comparison of Eqs. (3.16) and (3.20) shows that for the same bending moment, the stress obtained by using Eq. (3.20) from plastic analysis is 50% less than that obtained by using Eq. (3.16) for

elastic analysis. In other words, there is a stress reduction factor of 1.5 when using plastic analysis

compared to elastic analysis. This fact is utilized by the ASME code as well as other international

codes to increase the allowable elastic stress values by a factor of 1.5 when analyzing beams of rectangular cross-section, bending of plates, and local bending of shells. The ASME code uses a conservative shape factor of 1.25 for rectangular cross-sections in the creep range.

TABLE 3.2

VARIATION IN SHAPE FACTOR FOR RECTANGULAR

CROSS-SECTIONS WITH RESPECT TO n

n

SF

ASM_Jawad_Ch03.indd

1

1

67

2.5

1.25

5

1.36

10

1.43

MTC

1.50

02/24/2009 04:59AM

68

Chapter 3

FIG. 3.3

FROM BEEDLE (BEEDLE, 1958), REPRINTED WITH PERMISSION

3.3.2

Circular Cross-Sections

The relationship between maximum stress and moment for pipes of circular cross-section in the

elastic range is

V

MRo I

(3.21)

where

I = moment of inertia of a tubular cross-section = (Ro4 - Ri4)/4

Ri = inside radius of beams

Ro = outside radius of beams

ASM_Jawad_Ch03.indd

68

MTC

02/24/2009 04:59AM

Members in Bending 69

TABLE 3.3

VARIATION IN SHAPE FACTOR FOR HOLLOW

CIRCULAR CROSS-SECTIONS WITH RESPECT TO n

n

SF

1

1

2.5

1.14

5

1.20

10

1.243

1.27

The relationship between maximum stress and moment in the creep range is obtained from Eq.

(3.14) as

V

MRo In

(3.22)

or

V

1

SF

MRo

I

(3.23)

SF

1 R i Ro 3 1 n

*>1 12n@

1 Ri Ro 4

*>15 12n@

>3 1n@ S 1 2

(3.24)

For n = 1 the shape factor is 1.0, and Eq. (3.21) for elastic analysis and Eq. (3.23) for creep analysis become the same. Variation in SF with respect to n is shown in Table 3.3 for hollow thin circular

cross-sections (Ri >> t).

3.4

DEFLECTION OF BEAMS

The approximate relationship between the deection of a beam and the radius of curvature, , is

given by the mathematical equation

d2w

dx 2

1

U

(3.25)

where

w = deection of beam

Combining this equation with the stationary stress Eq. (3.2), as well as Eqs. (3.3) and (3.8) yields

the following expression for the deection of a beam

d2w

dx 2

K M nInn

(3.26)

where

In = moment of inertia as dened by Eq. (3.7)

M = applied moment on beam

Equation (3.26) is non-linear for all values of n, except n = 1.0. Thus, its application is very cumbersome for all but the simplest cases. Some of these simple cases (Faupel, 1981) are shown below.

Cantilever beam (Fig. 3.4a) with an end load F

ASM_Jawad_Ch03.indd

69

MTC

02/24/2009 04:59AM

70

Chapter 3

FIG. 3.4

CANTILEVER BEAM

GA

K F n Ln 2

K F L

TA

(3.27)

Inn n 2

I nnn

n1

(3.28)

1

GA

K q n L2n 2

(3.29)

I nn 2 n 2n 2

K q n L2n 1

TA

(3.30)

Inn 2 n 2n 1

n

GA

qA =

K MA L

(3.31)

2I nn

-K MAn L

(3.32)

I nn

The following example illustrates the application of Eq. (3.26) to a two-span piping system.

Example 3.2

Determine the maximum bending moments in the piping system shown in Fig. 3.5a. Assume a

uniform load q = 0.1 kips/ft due to piping weight and contents and L = 20 ft. Let n = 1 and 4, and

compare the results with plastic analysis.

ASM_Jawad_Ch03.indd

70

MTC

02/24/2009 04:59AM

Members in Bending 71

FIG. 3.5

TWO-SPAN BEAM

Solution

Due to symmetry, the structure is reduced to that shown in Fig. 3.5b. The bending moment in the

beam is

M

RAx qx 22

(a)

w

(b)

n=1

The value of RA is obtained by solving Eq. (b) for the three boundary conditions

w

0 at x

0 L and w

0 at x

L

This gives

RA

0375 qL and M

(c)

n=4

The value of RA is obtained by solving Eq. (b) for the three boundary conditions

w

ASM_Jawad_Ch03.indd

71

0 at x

0 L and w

0 at x

MTC

02/24/2009 04:59AM

72

Chapter 3

TABLE 3.4

SOLUTION OF EQS. (c) AND (d) IN EXAMPLE 3.2

Moment

Maximum negative

moment at point B

Maximum positive

moment at point D

Location of maximum

positive moment

qL2/14.22

x = 0.375L

qL2/8

Elastic analysis

(n = 1.0)

Creep analysis

(n = 4.0)

Plastic analysis

qL /10

qL /12.5

x = 0.40L

qL2/11.63

qL2/11.63

x = 0.414L

The expression for the reaction obtained from the boundary conditions is a fourth-order equation.

A numerical solution gives

RA

040qL and M

040qLx qx 22

(d)

Solutions of Eqs. (c) and (d) are shown in Table 3.4. When n = 1, Eq. (c) yields the moment expression for an elastic analysis. It has a maximum negative value at support B of MB = qL2/8 and a

maximum positive value at x = 0.375L of M = qL2/14.22. For n = 4.0, the maximum negative value

at support B is obtained from Eq. (d) as MB = qL2/10 and a maximum positive value at x = 0.40L of

M = qL2/12.5. Thus, for n = 4.0, the moments in the beam have redistributed where the negative and

positive moments are closer in magnitude. It is of interest to note that a plastic analysis for this structure gives a bending moment with a maximum negative value at support B of MB = qL2/11.63 and a

maximum positive value at x = 0.414L of M = qL2/11.63. Thus, the stationary creep analysis of this

structure can be approximated by performing the much easier plastic analysis even when the values of

n are relatively low. One method of plastic analysis is illustrated in the next section.

The stationary stress and plastic stress analyses discussed in the above example are fairly tedious to

perform for complex piping loops. The elastic analysis, however, gives conservative values as shown

in Table 3.4 and is much easier to perform. Accordingly, the ANSI B31.1 (Power Piping) and B31.3

(Process Piping) allow elastic analysis of piping loops in high-temperature applications. The allowable

stress values at the high temperatures are based on creep and rupture criteria. This elastic analysis and

some of the requirements of B31.3 are discussed next.

3.5

3.5.1

Introduction

A piping system is one of the more complex structural congurations a component designer or

analyst is likely to encounter. Although the straight sections are relatively simple, the connecting components, principally elbows and branch connections, have complex stress distributions even under

simple loading conditions. In addition, the piping system frequently contains other components such

as anged joints and valves. Added to this complexity is the fact that all components of the piping

system interact with each other such that the loading on any single component is a function of the

response of the whole piping system to applied loading conditions.

There are many references for the design and evaluation of piping systems and it is not the purpose

of this section to replicate those sources. Rather, the purpose is to (1) acquaint the designer/analyst

with the overall approach to piping design as embodied in B31.1 Power Piping and B31.3 Process

Piping, (2) discuss some of the issues impacting piping design in the creep regime, and (3) provide a

ASM_Jawad_Ch03.indd

72

MTC

02/24/2009 04:59AM

Members in Bending 73

hand evaluation of a simple pipeline to illustrate the application of B31.3. Although they differ in some

details, the overall approach in B31.1 and B31.3 is quite similar. The following discussion is based

largely on Chapters 16 and 17 by Becht in the Companion Guide to the ASME Boiler and Pressure

Vessel Code (Rao, 2002) and Chapter 38 by Rodabaugh in the same reference.

3.5.2

There are three general categories of design requirements in B31.1 and B31.3: pressure design, sustained and occasional loading, and thermal expansion.

3.5.2.1 Pressure Design. There are four basic approaches to pressure design: (1) components for

which there are tabulated pressure/temperature ratings in their applicable standard; (2) components

that are specied to have the same pressure rating as the attached piping; (3) components such as

straight pipe and branch connections for which design equations are provided to determine minimum

required thickness; and (4) non-standard components for which experimental methods, for example,

may be used. The design equations for minimum wall thickness are nominally based on Lames hoop

stress equations, which take into account the variation of pressure-induced stress through the wall

thickness. The allowable stress for pressure loading is based on criteria similar to those used for

Sections I and VIII Div 1. In the creep regime, this is the lower of 2/3 of the average creep rupture

strength in 100,000 hours, 80% of the minimum creep strength or the average stress to cause a minimum creep rate of 0.01% per 1000 hours.

3.5.2.2 Sustained and Occasional Loading. Whereas the pressure stress criteria are nominally

based on hoop stress, the sustained and occasional load limits are based on longitudinal bending

stresses. The limit for sustained loading (e.g., pressure and weight) is the same as for pressure design.

However, because the longitudinal stress due to pressure is nominally half the hoop stress, there is a

margin left for longitudinal bending and pressure stresses. The limit for occasional loading, pressure,

and weight plus seismic and wind, is a factor that is higher based on the loading duration.

Both B31.1 and B31.3 specify the use of 0.75 times the basic factor stress intensication factors for

sustained and occasional loading; however, Becht (in Rao, 2002) notes that some piping analysis computer programs use the full intensication factor for the sustained and occasional load limit. Considering also stress redistribution effects due to creep as discussed earlier in this chapter, it is recommended

that the full intensication factor be used in evaluation of sustained and occasional load stresses for

B31.1 and B31.3 applications in the creep regime.

3.5.2.3 Thermal Expansion. Not all piping systems require an analytical evaluation of piping

system stresses due to restrained thermal expansion. Guidelines for the exceptions are provided in

B31.1 and B31.3. When an analytical evaluation is made, it is based on elastic analysis using classical

global stiffness matrices based on superposition principals. However, the exibility of piping components such as elbows and miter joints is modied by exibility factors that approximate the additional

exibility of these components as compared to a straight pipe with the same length as the centerline length of the curved pipe. Formulas for these exibility factors are provided in the respective

codes.

In addition to enhanced exibility, stress levels in components such as elbows and branch connections are also higher than would be the case for a straight pipe under the same loads. To account for

these higher stresses, stress intensication factors are also provided. Component stresses are then

calculated from the equation

2

i i Mi i o Mo

Sb

ASM_Jawad_Ch03.indd

73

1 2

MTC

(3.33)

02/24/2009 04:59AM

74

Chapter 3

where

ii = in-plane stress intensication factor (see Table 3.5)

io = out-plane stress intensication factor (see Table 3.5)

Mi = in-plane bending moment

Mo = out-plane bending moment

Sb = resultant bending stress

Z = section modulus of pipe

The allowable stress range for restrained thermal expansion is designed to achieve shakedown in a

few number of cycles. The background on shakedown concepts is discussed in more detail in Section

1.6.3. The basic limit (Severud, 1975 and 1980) for allowable range of restrained thermal expansion

stresses in B31.1 and 31.3 is given by the equation

SA

f 125Sc 025Sh

(3.34)

where

SA = allowable displacement stress range

Sc = allowable stress at the cold end of the thermal cycle

Sh = allowable stress at the hot end of the thermal cycle

f = stress range reduction factor used to account for fatigue effects when the number of equivalent

full-range thermal cycles exceeds 7000 (about once a day for 20 years)

The above equation does not give credit for the unused portion of the allowable longitudinal stress,

SL, discussed in Section 3.5.2.2. If such credit is taken, Eq. (3.34) becomes

SA

f >125Sc Sh SL @

(3.35)

Although Eq. (3.35) provides a somewhat higher allowable stress, the difference becomes small as

the temperature moves into the creep regime and the allowable stress is dependent on creep properties. Equation (3.34) has the advantage that thermal expansion stresses may be evaluated separately

from the local values of longitudinal stresses, SL.

TABLE 3.5

FLEXIBILITY AND STRESS INTENSIFICATION FACTORS (ASME, B31.3-2004)

Stress Intensication

Description

Welding elbow

or pipe bend

Closely spaced

miter bend

s < r2 (1 + tan )

ASM_Jawad_Ch03.indd

Flexibility

factor,

k

165

h

152

h56

74

Out-ofplane,

Io

In-plane,

Ii

075

h23

09

h23

09

h23

09

h23

Flexibility

characteristic,

h

Sketch

T R1

r22

cot T

2

sT

r22

MTC

02/24/2009 04:59AM

Members in Bending 75

In the above equations, the stress range reduction factor, f, is based on a fatigue curve for buttwelded carbon steel pipe developed by Markl (1960).

3.5.3

Creep Effects

There are several areas where creep effects play an important role in piping design, among them are

weld strength reduction factors, elastic follow-up, and the effects of creep on cyclic life.

3.5.3.1 Weld Strength Reduction Factors. Weld strength reduction factors were introduced in

the 2004 edition of B31.3 to account for the reduced creep rupture strength of weld metal as compared

to base metal. Although the designer may use factors based on specic data, a general factor is provided that is applicable to all materials in lieu of specic weld metal data. This factor varies linearly

from 1.0 at 950F to 0.5 at 1500F.

3.5.3.2 Elastic Follow-Up. The use of elastic analysis procedures to determine the stress/strain

distribution due to restrained thermal expansion is based on the assumption that the piping system

is balanced without localize high stress areas. Examples of piping congurations that are unbalanced

are provided in B31.3, but there are no quantitative criteria provided. An unbalanced system operating in the creep regime can be particularly susceptible to elastic follow-up considerations as discussed

in Section 1.6.1. Under severe follow-up conditions with localized stress/strain concentrations, the

resultant localized stress due to restrained thermal expansion may not relax, thus behaving more like a

sustained load. Under these circumstances, i.e., a signicantly unbalanced system, it may be appropriate to treat these stresses similar to a sustained load with an allowable stress of Sh.

3.5.3.3 Cyclic Life Degradation. The tests on which the f factors are based were performed on

carbon and stainless steel components below the creep regime. Operating in the creep regime can

have several deleterious effects. First, the basic continuous cycling fatigue curve will be lower at higher

temperatures. Second, as previously discussed in Section 1.6.4, there is a hold-time effect on cyclic

life due to the accumulation of creep damage as restrained thermal expansion stresses relax. And,

third, due to Neuber-like effects, the strain at structural discontinuities will be greater than predicted

by elastic analyses. Taking direct account of these phenomena goes well beyond the scope of normal

piping analyses. There is, however, a conservative approximation included in Code Case N-253 for

Class 2 and 3 nuclear components at elevated temperature. The changes in the f factor as described in

Appendix B of Code Case N-253 are an indication of the potential signicance of the effects of creep

on cyclic life. For carbon steel, instead of a reduction starting at 7000 cycles below the creep regime,

at 750F the reduction starts at 50 cycles and decreases to 5 cycles at 900F. For 304 SS, the reduction

starts at 850F, reducing to 50 cycles at 1000F and 5 cycles at 1200F.

The above discussions should not be taken as direct recommendations for reduction in allowable

stress levels for B31.1 and B31.3 in the creep regime. Application of these criteria has, after all, led to

a long history of successful service experience. It is, however, a cautionary recommendation against

pushing the criteria to their limits particularly when dealing with unfavorable geometries and a

large number of service cycles.

3.6

STRESS ANALYSIS

One method of analyzing a piping system in the creep region is to perform an elastic analysis. A

three-dimensional frame analysis is usually performed for piping loops. Such an analysis is beyond

the scope of this book. In this chapter, however, the moments Mi are obtained from a simple twodimensional structural analysis to demonstrate the procedure. The general elastic relationship between applied joint loads and forces, Fig. 3.6, in a frame is given by

>P @

ASM_Jawad_Ch03.indd

75

>K @> X @

MTC

(3.36)

02/24/2009 04:59AM

76

Chapter 3

FIG. 3.6

SIGN CONVENTION FOR BEAMS

where

[P] = applied loads at frame joints

[K ] = global stiffness matrix of frame given in Eq. (3.37)

[X ] = deformation of frame joints

Figure 3.6 shows a typical member in bending. Positive direction of forces, applied loads, joint

deections, and angles are shown in Fig. 3.6. Member 1, shown in the gure, starts at node 1 and

ends at node 2. At node 1, forces P1, P2, and moment M1, or deections X1, X2, and rotation 1, may

be applied as shown in the gure. Similarly, forces P1, P2, and moment M1, or deections X1, X2, and

rotation 1, can also be applied at node 2. The local member rotation angle with respect to global

coordinate system of the member is measured positive counterclockwise from the positive x direction

at the beginning of the member. Tensile force is as shown in the gure.

The global stiffness matrix of the frame is assembled from the individual frame member stiffnesses

[k ]. The stiffness matrix, [k ], of any member in a frame (Weaver and Gere, 1990) is expressed as

k 14 k 15 k 16

k 24 k 25 k 26

k 44 k 45 k 46

ASM_Jawad_Ch03.indd

76

k 34 k 35 k 36

k 55 k 56

k 66

(3.37)

>k @

k 11 k 12 k 13

k 22 k 21

k 13

symmetric

MTC

02/24/2009 04:59AM

Members in Bending 77

where

k 11 = 4EI/L

k 12 = (6EI/L2)sin

2

k 13 = -(6EI/L )cos , k 14 = 2EI/L

k 16 = -k 13

k 15 = -k 12,

k 22 = (EA/L)cos2 + (12EI/L3)sin2

k 23 = (EA/L)sin cos - (12EI/L3)sin cos

k 25 = -k 22,

k 26 = -k 23

k 24 = k 12,

k 33 = (EA/L)sin2 + (12EI/L3)cos2

k 35 = k 26,

k 36 = -k 33

k 34 = k 13,

k 44 = k 11,

k 45 = - k 12,

k 46 = -k 13

k 55 = k 22,

k 56 = k 23,

k 66 = k 33

Equation (3.37) is a 6 6 matrix because there is a total of six degrees of freedom at the ends of the

member. They are X1 through X6 or P1 through P6. The stiffness matrix of the frame is assembled by

combining the stiffness matrices of all the individual members into one matrix, called the global matrix. Once the global stiffness matrix [K ] is developed, Eq. (3.36) can be solved for the unknown joint

deections [X ]. This assumes that the load matrix [P] at the nodal points is known.

>X @

(3.38)

The forces in the individual members of the frame are then calculated from the equation

>F @

>D@>Xm @

(3.39)

where

[F] = force in member

[D] = force-deection matrix of a member

[Xm] = matrix for deection at ends of individual members

The matrix [D] is dened as

D 15

D 16

D 23 D 24

D 25

D 26

D 33 D 34

D 35

D 36

D 12

D 13

D 22

D 32

(3.40)

D 21

D 31

where

D12 = -(AE/L)cos ,

D16 = -D13,

D23 = -(6IE/L2)cos ,

D26 = -D23, D31 = D24,

D35 = D25,

D36 = D26

D13 = -(AE/L)sin ,

D21 = 4EI/L,

D24 = 2EI/L,

D32 = D22, D33 = D23,

D15 = -D12

D22 = (6IE/L2)sin

D25 = -D22

D34 = D21

The application of these equations to a simple piping loop is illustrated in the following example.

Example 3.3

The pipe loop shown in Fig. 3.7a has a 4-in. STD weight pipe with an OD = 4.5 in., thickness =

0.237 in., and weight = 10.79 lb/ft. The pipe carries uid with density of 62.4 lb/ft3. Determine the

maximum stress due to pipe and uid weight using elastic analysis. The pipe hanger CE prevents vertical deection at that location. Let the modulus of elasticity = 27 106 psi. The elbows at points B and

C have a bend radius of 6.0 in.

ASM_Jawad_Ch03.indd

77

MTC

02/24/2009 04:59AM

78

Chapter 3

FIG. 3.7

PIPE LOOP

Solution

(a) Elastic analysis

From Fig. 3.7b,

Member AB:

ASM_Jawad_Ch03.indd

3.175 in.2 ,

0 deg.,

78

sin D

7.210 in.4 , L

0.0,

cos D

240 in.

1.0

MTC

02/24/2009 04:59AM

Members in Bending 79

1

3,244,639

2

0

357,080

3

-20,279

0

168.9916

4

1,622,319

0

-20,279

3,244,639

Symmetric

5

0

-3,577,080

0

0

357,080

6

20,279

0

-168.9916

20,279

0

168.9916

1

2

3

4

5

6

9

0

0

4

5

-357,080

0

0

357,080

6

7

8

9

12

20,279

0

-168.9916

20,279

0

168.9916

7

8

9

10

11

12

Member BC:

A

3.174 in. 2,

90 deg,

sin D

7.210 in.4, L

240 in.

cos D

1.0,

0.0

4

3,244,639

5

20,279

168.9916

6

0

0

7

1,622,319

20,279

357,080

0

3,244,639

Symmetric

8

-20,279

-168.9916

0

-20,279

168.9916

Member CD:

A

3174 in. 2

0 deg

sin D

7210 in4

cos D

00

240 in

10

7

3,244,639

8

0

357,080

9

-20,279

0

168.9916

Symmetric

10

1,622,319

0

-20,279

3,244,639

11

0

-3,577,080

0

0

357,080

From Figs. 3.6 and 3.7, it is seen that the degrees of freedom 1, 2, 3, 9, 10, 11, and 12 must be set to

zero because the pipe loop is xed at these nodal points. This reduces the stiffness matrix k AB to a 3 3

matrix corresponding to displacements 4, 5, and 6. It also reduces stiffness matrix k BC to a 5 5 matrix

corresponding to displacements 4, 5, 6, 7, and 8, and matrix k CD to a 2 2 matrix corresponding to

displacements 7 and 8. The total global matrix is then a 5 5 matrix with magnitude Kglobal =

4

6,489,278

5

20,279

357,249

6

20,279

0

357,249

7

1,622,319

20,279

0

6,489,278

Symmetric

8

-20,279

-169.0

0

-20,279

357,249

4

5

6

7

8

ASM_Jawad_Ch03.indd

79

MTC

02/24/2009 04:59AM

80

Chapter 3

xed end moments 16.31(20)2/12

reactions

543.7 ft- lb

16.31 lb/ft

6524 in.- lb

16.31(20)/2 163.1 lb

6524

1631

6524

>F @

X4

X5

X6

X7

X9

4

13.3920 u 10

5

38.0525 u 10

13.4015 u 10 4

in.

in.; downward deection

rad; clockwise rotation

in.

1. Member 1

F

MA

MB

lb

in. - lb =

in. - lb

0

- 8700

2172

0

- 725

181

lb

ft - lb

ft - lb

135 .9

- 181

181

lb

ft - lb

ft - lb

2. Member 2

F

MB

MC

135.9

- 2172

2172

lb

in. - lb =

in. - lb

3. Member 3

F

MC

MD

= - 2172

8700

lb

in. - lb =

in. - lb

0

- 181

725

lb

ft - lb

ft - lb

Figure 3.7c shows the nal forces and bending moments in members AB, BC, and CD.

Maximum bending stress at the supports

S

8700225

721

2715 psi

This value must be less than the allowable stress at the given temperature.

ASM_Jawad_Ch03.indd

80

MTC

02/24/2009 04:59AM

Members in Bending 81

From Table 3.5,

h

0.237(6)/(2.13) 2

2/3

ii

0.9/ 0.313

181(1.952)

0.313

1.952

353.3 ft - lb

353312225

721

1323 psi

This value must be less than the allowable stress at the given temperature.

3.6.1

Commercial Programs

Numerous computer programs are available for determining moments and forces in piping loops

such as AutoPIPE, CAEPIPE, CEASAR II, and PipePak. These sophisticated programs take into

consideration such items as pressure, thermal expansion, dead weight, uid weight, support movements, and hanger locations. Most of the programs also take into account the reduction in moments

and forces in the piping loop due to elbow radii and other curved members. They also check stresses

in accordance with various codes and standards such as B31.1 and B31.3.

Example 3.4

The nal forces, moments, and stress calculated in Example 3.3 were compared with results obtained from the commercial piping program AutoPIPE. The program was run with the same loads

and geometry of Example 3.3. The nal forces and bending moments calculated by the program are

shown in Fig. 3.8. A comparison between the bending moments obtained manually (Fig. 3.7c), and

those obtained from AutoPIPE (Fig. 3.8) show very good agreement at the supports and 15% variance

at the elbows, with the manual method being on the conservative side. The 15% difference is because

the manual method assumes a sharp corner at points B and C, whereas AutoPIPE takes the elbow

curvature into consideration.

The stress at the supports and elbows calculated by AutoPipe are

Stress at supports = 2753 psi

Stress at elbows = 1169 psi

The stress at the elbow takes into consideration the stress intensication factor calculated from

B31.1.

3.7

The reference stress method, discussed in Section 1.5.3.2, is based on the following equation to

determine the equivalent creep stress in a structure

R = (P/PL)y

(3.41)

where

P = applied load on the structure

PL = the limit value of applied load based on plastic analysis

ASM_Jawad_Ch03.indd

81

MTC

02/24/2009 04:59AM

82

Chapter 3

FIG. 3.8

MOMENT AND FORCE OBTAINED FROM AUTOPIPE

R = reference stress

y = yield stress of the material

There are many methods of performing plastic analysis in a piping system. One such method is to

use elastic equations with incremental applied load conditions until one of the joints reaches its plastic limit. Then the structure is modied and a second run is performed until a second joint reaches

its limit. The analysis continues until the structure becomes unstable at which point the analysis is

terminated and the maximum load carrying capacity of the structure is obtained. Another method,

for simple piping systems, is to use plastic structural analysis. This is demonstrated in the following

example.

Example 3.5

Solve Example 3.3 using plastic structural analysis. Use a shape factor of 1.27 from Table 3.3 and

assume y = 20,000 psi.

Solution

The plastic moments in the pipe loop are shown in Fig. 3.9. Due to symmetry, member AB will be

used.

of internal energy = of external energy

Mp T Mp 2 T Mp T

ASM_Jawad_Ch03.indd

82

P LL G 2

MTC

(a)

02/24/2009 04:59AM

Members in Bending 83

FIG. 3.9

PLASTIC MOMENTS IN A PIPE LOOP

Let tan /(L/2). Then Eq. (a) becomes

PL L216

Mp

Vy

Mp R 2

I SF

PL

2728 lb/ft

VR

This value is substantially lower than the stress value of 2716 psi obtained from elastic analysis in

Example 3.3.

3.8

CIRCULAR PLATES

The differential equations for the bending of circular plates in the creep regime have no closedform solution. Numerical solutions obtained by Odqvist (1966), Kraus (1980), and other authors have

shown that results of stationary stress solution of circular plate equations in the creep range approach

the results obtained by plastic theory as the value of n increases. For n = 1, the elastic and creep results

are the same.

Table 3.6 shows various values of the maximum bending moment in a circular plate using elastic

and plastic analysis. For a simply supported plate, the maximum bending moment due to lateral pressure and based on elastic analysis is

ASM_Jawad_Ch03.indd

83

MTC

02/24/2009 04:59AM

84

Chapter 3

TABLE 3.6

MAXIMUM MOMENT IN A SOLID CIRCULAR PLATE1 (FROM JAWAD [2004])

Simply supported edge2

Loading condition

Uniform load, P

(elastic solution)

Uniform load, P

(plastic solution)

Center load, P

(plastic solution)

Fixed edge3

PRo /4.85

PRo2/8

PRo2/6

PRo2/12

F/(2)

F/(4)

Maximum moment is at the center of plate.

3

Maximum moment is at the edge of the plate.

F, concentrated load at center of plate; P, uniform applied pressure; Ro, radius of plate.

2

PR 2485

(3.42)

where

M = bending moment

P = applied pressure

Ro = radius of plate

The required thickness is obtained from the equation S = 6M/t2 as

Ro >1237 PS @1 2

(3.43)

where

S = allowable stress

t = thickness

Equation (3.43) can be rewritten in a slightly different form by substituting SE for S and d for 2Ro.

The result is

d>CPSEo @1 2

(3.44)

where

d = diameter of the plate

Eo = joint efciency

C = constant (0.309 for simply supported plate 0.188 for xed plate)

The ASME Sections I and VIII-1 codes use Eq. (3.44) for a variety of unstayed at heads. Numerous values of C are given for the various head congurations and shell attachment details. Sections I

and VIII-1 use Eq. (3.44) for all temperatures ranges including those in the creep range.

It is of interest to note that a plastic analysis, performed as a limit case for creep analysis, results

in a maximum moment for a simply supported case as M = PR2/6 as shown in Table 3.6. From the

expression S = 6M/t2 and using a shape factor of 1.25 from Table 3.2 gives

t

ASM_Jawad_Ch03.indd

84

d>02PS @12

MTC

(3.45)

02/24/2009 04:59AM

Members in Bending 85

A comparison of Eqs. (3.44) and (3.45) shows a thickness reduction of 20% is achieved by performing plastic analysis compared to elastic analysis for the same allowable stress.

Problems

3.1 The 12-in. diameter Standard Schedule pipe loop is as shown in the gure. The pipe loop and

the vessels undergo a temperature increase from ambient to 1000F. Calculate the thermal

stress in the pipe loop and compare it with the allowable stress values. The allowable stress at

ambient temperature is 17,100 psi and at 1000F it is 8000 psi. Let f = 1.0, E = 20.4 106 psi,

= 8.2 10-6 in./in.-F, A = 14.6 in.2, I = 279 in.4, t = 0.375 in.

ASM_Jawad_Ch03.indd

85

MTC

02/24/2009 04:59AM

ASM_Jawad_Ch04.indd

86

MTC

02/24/2009 04:07AM

CHAPTER

4

ANALYSIS OF ASME PRESSURE

VESSEL COMPONENTS:

LOAD-CONTROLLED LIMITS

4.1

INTRODUCTION

The ability of pressure vessel shells to perform properly in the creep range depends on many factors

such as their stress level, material properties, temperature range, as well as operating temperature and

pressure cycles. American Society of Mechanical Engineers (ASME) Sections I and VIII-1, which are

based on design-by-rules rather than design by analysis, allow components to be designed elastically

at elevated temperature service as long as the allowable stress is based on creep-rupture data obtained

from Section II-D. In many cases, however, additional analyses are required to ascertain the adequacy

of such components at elevated temperatures. The effect of creep on long-time exposure of various

components in power boilers to high temperatures is of interest. The effect of continuous startup and

shutdown of heat recovery steam generators (HRSG) on the life expectancy at elevated temperatures

is one of the design issues. Another safety concern is the effect of long-time exposure of equipment

used in the petrochemical industry to high temperatures.

These additional analyses include evaluation of bending and thermal stresses as well as creep and

fatigue. The sophistication of the analyses could range from a simple elastic solution to a more complicated plastic or a very complex creep analysis. The preference of most engineers is to perform an

elastic analysis because of its simplicity, expediency, and cost effectiveness. The approximate elastic

analysis is not as accurate as plastic or creep analysis, which reects more accurately the stress-strain

interaction of the material. However, elastic analysis is sufciently accurate for most design applications. The ASMEs elastic procedure compensates for this approximation by separating the stresses

obtained from the elastic analysis into various stress categories in order to simulate, as close as practically possible, the more accurate plastic or creep analysis.

The 2007 edition of ASME Section VIII-2 covers temperatures in the creep regime above the previous limits of 700F and 800F for ferritic and austenitic materials, respectively. Section VIII-2 also

requires either meeting the requirements for exemption from fatigue analysis, or, if that requirement is

not satised, meeting the requirements for fatigue analysis. However, above the 700/800F limit, the

only available option is to satisfy the exemption from fatigue analysis requirements because the fatigue

curves required for a full fatigue analysis are limited to 700F and 800F.

The additional analyses for Section I and VIII-1 vessels at elevated temperatures are normally

performed with the aid of VIII-2 and III-NH of the ASME code as well as other international codes.

Section VIII-2, although generally limited to temperatures below the creep range, contains detailed

87

ASM_Jawad_Ch04.indd

87

MTC

02/24/2009 04:07AM

88

Chapter 4

rules for establishing various stress categories needed in the creep analysis of pressure parts. Section

III-NH contains rules for creep analysis of various stress components. And although III-NH is intended for nuclear service, many designers use the rules for non-nuclear applications as well. The

proper integration of the rules of VIII-2 and III-NH in analyzing boiler and pressure vessel components in accordance with I and VIII-1 at elevated temperatures is one of the subjects discussed in this

chapter. It should be noted that creep analysis is only applicable when the material is ductile. Nonductile materials such as cast iron or steels that are embrittled during operation cannot be analyzed by

the criteria of VIII-2 and III-NH because the results can lead to unsafe performance.

ASME nuclear Sections III-NB, NC, and ND apply to equipment operating at temperatures below

the creep range. At elevated temperatures in the creep regime, the requirements of III-NH are used

for class 1 components and the requirements of Code Case N-253 are used for class 2 and 3 components.

The designer of equipment operating in the creep range must rst comply with the applicable ASME

code of construction such as I, VIII-1, or VIII-2, and then meet the requirements of III-NH for creep

analysis. However, the allowable stress values, at a given temperature, are different for each of these

code sections at both below and above the creep temperature range. This difference is due, in part,

to the differences in safety factors established by various codes. This is demonstrated in Table 4.1 for

2.25Cr-1Mo steel and 304 stainless steel. The table shows a substantial difference in allowable stress

from one code section to the other. Accordingly, and in order to be consistent, the following stress

criteria will be used throughout this chapter as well as Chapters 5 and 6:

Stress criteria in Chapters 4, 5, and 6 are

The allowable stress, S, from the applicable I, VIII-1, or VIII-2 code sections will be used for

design purposes.

The allowable stress S from VIII-2 will be used for analyzing components below the creep

range.

The allowable stresses Sm, Smt, and St from III-NH will be used for analyzing components in

the creep range.

The procedure for creep analysis in III-NH consists of calculating rst a trial thickness of a component. The component is then analyzed for load-controlled stress limits and then for strain-controlled

stress limits. In this chapter, the load-controlled stress limits are discussed, whereas the strain-controlled stress limits are described in Chapter 5.

TABLE 4.1

ALLOWABLE DESIGN CONDITION STRESS VALUES FOR TWO MATERIALS AT ROOM

TEMPERATURE AND 1000F

Material

Temperature

(F)

I

S

VIII-1

S

VIII-2

S

III-NB

Sm

III-NH

So

SA 387 22CL11

SA 387 22CL1

SA240-3044

SA240-304

RT

1000

RT

1000

17.1

8.03

20.0

14.0

17.1

8.0

20.0

14.0

20.0

8.0

20.0

14.0

20.0

NA

20.0

NA

NA2

8.0

NA2

11.1

Tensile stress = 60 ksi at room temperature (RT) and 53.9 ksi at 1000F. Yield stress = 30 ksi at RT and 23.7 ksi at

1000F.

2

Not applicable.

3

Stress values in italics are controlled by creep and rupture criteria.

4

Tensile stress = 75 ksi at RT and 57.4 ksi at 1000F. Yield stress = 30 ksi at RT and 15.5 ksi at 1000F.

ASM_Jawad_Ch04.indd

88

MTC

02/24/2009 04:07AM

4.2

DESIGN THICKNESS

The required shell thickness for Section I and VIII vessels at any temperature is obtained from the

corresponding equations given in these two sections. When an additional stress analysis is specied or

required at elevated temperatures, then the procedure established in III-NH for creep analysis (Fig.

4.1) is used as a guideline. The procedure is based on rst calculating a thickness of the component.

Equations and allowable stress from I or VIII as applicable

With time dependent allowable stresses

Yes

No

Redesign

No

Yes

operating conditions in accordance with chapter

5.

Yes

No

inelastic analysis in accordance with chapter 5.

Analysis is complete

FIG. 4.1

SEQUENCE IN CREEP ANALYSIS FOR DESIGN AND OPERATING CONDITIONS

ASM_Jawad_Ch04.indd

89

MTC

02/24/2009 04:07AM

90

Chapter 4

The thickness is based on design conditions and is obtained from applicable design equations in Section I or VIII. Stresses due to operating conditions are then calculated and maintained below specied

allowable limits obtained from Section III-NH.

For illustration purposes, the following equations for cylindrical shells are given and will be used

throughout the next two chapters in solving various problems.

4.2.1

Section I

The design shell thickness when thickness does not exceed one-half the inside radius is given by

t

PDo

c

2SEo 2yP

(4.1)

PR i

c

SEo 1 yP

(4.2)

or

t

where

y = temperature coefcient as given in Table 4.2

c = corrosion allowance

The y factor in the above two equations was introduced in Section I in the mid 1950s (Winston et

al., 1954) to take into account the reduction of stress due to redistribution when the temperature is in

the creep region.

The design shell thickness when thickness exceeds one-half the inside radius is

t

>Z1 1 2 1@Ri

(4.3)

This equation is based on Lames Eq. (4.9), discussed later in this chapter. Section I recently added

a new shell design equation in Appendix A of Section I and is given by

t

D i e PSEo 1

(4.4)

2 c f

where

f = thickness factor for expanded tube ends

Equation (4.4) is based on limit analysis. It is applicable at low and well as high temperatures.

However, caution must be exercised when using this equation at elevated temperatures for thick-wall

TABLE 4.2

THE y COEFFICIENTS IN EQS. (4.1) AND (4.2) (ASME, I)

Temperature1

900 (480)

and below

950

(510)

1000

(540)

1050

(565)

1100

(595)

1150

(620)

1200

(650)

1250 (675)

and above

0.4

0.4

0.4

0.4

0.4

0.4

0.4

0.4

0.4

0.5

0.4

0.4

0.4

0.4

0.4

0.4

0.4

0.4

0.7

0.4

0.4

0.4

0.4

0.4

0.4

0.4

0.4

0.7

0.4

0.4

0.4

0.4

0.4

0.4

0.4

0.7

0.5

0.4

0.4

0.4

0.5

0.5

0.4

0.7

0.7

0.4

0.4

0.4

0.7

0.7

0.4

0.7

0.7

0.5

0.5

0.5

0.7

0.7

0.5

0.7

0.7

0.7

0.7

0.7

0.7

0.7

Ferritic

Austenitic

Alloy 800

800H, 800 HT

825

230 Alloy

N06045

N06690

Alloy 617

1

ASM_Jawad_Ch04.indd

90

MTC

02/24/2009 04:07AM

cylinders or material embrittled with aging because cracking on the surface will cause loss of load carrying capacity.

4.2.2

Section VIII

The design shell thickness in VIII-1 when thickness does not exceed one-half the inside radius or

pressure does not exceed 0.385SEo is given by

PR i

SEo 06P

(4.5)

c

The design shell thickness in VIII-1 when thickness exceeds one-half the inside radius or pressure

exceeds 0.385SEo is

Z12 1Ri c

(4.6)

The design shell thickness in VIII-2, part 4 (Design by Formula), is

t

D i ePSEo 1

(4.7)

Example 4.1

A pressure vessel is constructed of alloy 2.25Cr-1Mo steel and has in inside radius of 5 in. The

design temperature is 1000F, the design pressure is 3000 psi, and the allowable stress is 7800 psi.

The joint efciency, Eo, is equal to 1.0 and the corrosion allowance is zero. What is the required thickness

(a) In accordance with VIII-1?

(b) In accordance with I?

(c) In accordance with VIII-2?

Solution

(a) From Eq. (4.5)

3000 5

00

7800 10 06 3000

25 in.

t

3000 5

00

7800 10 1 07 3000

217 in.

t

235 in.

Notice that the thickness obtained from Section I is less than that obtained from Section VIII-1 due

to the y factor adjustment for elevated temperature. However, the obtained thicknesses from Sections

I and VIII-1 would be identical at lower temperatures, where the y factor in Eq. (4.2) is taken as 0.4.

ASM_Jawad_Ch04.indd

91

MTC

02/24/2009 04:07AM

92

Chapter 4

The thickness of a section I shell at 1000F obtained from Eq. (4.4), which has recently been added

to Section I, is 2.35 in. Thus, at this high temperature it is more economical to use the 2.17 in. obtained from Eq. (4.2). However, at temperatures below the creep value, Eqs. (4.2) and (4.4) yield

about the same results.

The thickness obtained from the VIII-2 equation is less than that obtained from the VIII-1 equation

for the same allowable stress. This is due to the stress redistribution taken into account in Eq. (4.7) as

a result of limit analysis.

All cylindrical shell equations in Sections I and VIII are based on Lames equation, or an approximation of it. Lames equation, which is based on elastic analysis, is expressed as

Ro R 2 1

(4.8)

J2

where

R = radius

= Ro/Ri

This equation has a maximum value at the inner surface where R = Ri of

V

J2 1

(4.9)

J2 1

2P J 2 1

(4.10)

Example 4.2

Find the inside and outside stress values in the VIII-1 and I vessels of Example 4.1 using Lames

Eqs. (4.9) and (4.10).

Solution

(a) For Section VIII-1 with t = 2.5 in.

J

755

15

1.52 + 1

s i = 3000

1.5 2 - 1

= 7800 psi

The outside stress is obtained from Eq. (4.10)

2300015 2 1

Vo

4800 psi

J

7175

1434

s i = 3000

1.434 2 + 1

1.434 2 - 1

= 8680 psi

ASM_Jawad_Ch04.indd

92

MTC

02/24/2009 04:07AM

This value, which is based on elastic analysis, is larger than the allowable stress at elevated temperatures. However, at a sustained elevated temperature, this stress relaxes to a lower value due to creep

(as discussed later in this chapter).

The outside stress is

s o = 2(3000)

1.4342 - 1

= 5680 psi

A plot of Eq. (4.9) for a thick shell is shown as line AB in Fig. 4.2. This line shows a non-linear

stress distribution across the thickness. The ASME Boiler and Pressure Vessel Code, Sections VIII-2

and III, as well as some other international codes, splits this non-linear calculated stress into different

components when an elastic stress analysis is performed. Methods of splitting this stress into various

categories and the denitions of these categories are discussed next.

4.3

STRESS CATEGORIES

The evaluation for the load-controlled stress limits is made in accordance with Fig. 4.1. The procedure for creep analysis in III-NH consists of rst calculating a trial thickness of the member and

then performing an elastic stress analysis to ascertain its adequacy in the creep regime. The stress in

the member is normally due to various operating conditions such as pressure, temperature, and other

loads. These loads may result in a complex stress prole across the thickness. This complex stress

prole is separated into various categories to simplify the elastic analysis. Each of the stress categories

is then compared to an allowable stress or equivalent strain limit. To demonstrate this procedure, it

is necessary to dene the various stress categories in accordance with the ASME code. ASME VIII-2

and III dene three different stress categories as shown in Fig. 4.3. They are referred to as the primary

stress, secondary stress, and peak stress. The ASME description of these stress categories follows.

4.3.1

Primary Stress

Primary stress is any normal or shear stress developed by an imposed loading, which is necessary to

satisfy the laws of equilibrium of external and internal forces and moments. The basic characteristic

FIG. 4.2

CIRCUMFERENTIAL STRESS DISTRIBUTION THROUGH THE THICKNESS OF

A CYLINDRICAL SHELL

ASM_Jawad_Ch04.indd

93

MTC

02/24/2009 04:07AM

94

Chapter 4

Stress Categories

Primary, Pm , PL , Pb

Secondary, Q

imposed loading which is

necessary to satisfy the laws

of equilibrium. The basic

characteristic of a primary

stress is that it is not self

limiting. Primary stress which

considerably exceeds the yield

strength will result in a failure

or at least in gross distortion.

A thermal stress is not

classified as a primary stress.

Peak, F

constraint of a structure.

Secondary stress is self limiting. Local yielding

and minor distortions can

satisfy the conditions

which cause the stress to

occur and failure from

one application of the

stress is not to be

expected.

cause any noticeable

distortion and is

objectionable only as

a possible source of a

creep failure, fatigue

crack, or brittle

fracture.

FIG. 4.3

STRESS CATEGORIES

of a primary stress is that it is not self-limiting. Primary stress is further subdivided into two categories

as shown in Fig. 4.4. The rst category is primary membrane stress, which could either be general Pm

or local PL, whereas the second category is primary bending stress, Pb.

4.3.1.1 General Primary Membrane Stress (Pm). This stress is so distributed in the structure

that no redistribution of load occurs as a result of yielding. An example is the stress in a circular cy-

Primary stress

Membrane

Bending, Pb ,

General, Pm

Local, PL

one which is so distributed in the

structure that no re-distribution of the

load occurs as a result of yielding.

small area such as a nozzle attachment or a

lug.

FIG. 4.4

PRIMARY STRESS

ASM_Jawad_Ch04.indd

94

MTC

02/24/2009 04:07AM

lindrical or spherical shell, away from discontinuities, due to internal pressure or to distributed live

loads.

4.3.1.2 Local Primary Membrane Stress (PL). Cases arise in which a membrane stress produced

by pressure or other mechanical loading and associated with a primary or a discontinuity effect produces excessive distortion in the transfer of load to other portions of the structure. Conservatism

requires that such a stress be classied as local primary membrane stress even though it shows some

characteristics of a secondary stress. Examples include the membrane stress in a shell produced by

external loads and moment at a permanent support or at a nozzle connection.

It is important to emphasize that the above denition includes pressure-induced loads as well as

those due to mechanical loading, and that the denition is equally applicable at any location in the

structure.

4.3.1.3 Primary Bending Stress (Pb). This stress is the variable component of normal stress in a

cross-section. An example is the bending stress in the central portion of a at head due to pressure.

4.3.2

Secondary Stress, Q

Secondary stress is a normal stress or a shear stress developed by the constraint of adjacent material

or by self-constraint of the structure, and thus it is normally associated with deformation-controlled

quantity at elevated temperatures. The basic characteristic of a secondary stress is that it is self-limiting. Local yielding and minor distortions can satisfy the conditions that cause the stress to occur and

failure from one application of the stress is not to be expected. Examples of secondary stresses are

bending stress at a gross structural discontinuity, bending stress due to a linear radial thermal strain

prole through the thickness of section, and stress produced by an axial temperature distribution in a

cylindrical shell.

The designer must keep in mind that, in some cases, NH considers thermally induced stresses

(NH-3213(c)) as primary. Also, pressure-induced discontinuity stresses are sometimes considered

primary (T-1331(d)).

4.3.3

Peak Stress, F

Peak stress is that increment of stress that is additive to the primary plus secondary stresses by reason of local discontinuities of local thermal stress including effects of stress concentrations. The basic

characteristic of a peak stress is that it does not cause any noticeable distortion and is objectionable

only as a possible source of a fatigue crack or a brittle fracture, and, at elevated temperatures, as a

possible source of localized rupture or creep fatigue failure. Some examples of peak stress are thermal

stress at local discontinuities and cladding, and stress at a local discontinuity.

4.3.4

Separation of Stresses

A complex stress pattern in a cross-section must be separated into primary, secondary, and peak

stress components in accordance with ASME. Table 4.3 lists the stress categories of some commonly

encountered load cases. The table shows that internal pressure for cylindrical shells must be divided

into primary membrane, Pm, and secondary, Q, stresses. Accordingly, the parabolic stress distribution given by Eq. (4.9) and shown as area ABCD in Fig. 4.2 must be decomposed into three parts as

shown in Fig. 4.5. The membrane stress CDEF is obtained by integrating area ABCD over the thickness and dividing by the total thickness. This results in the simple equation

ASM_Jawad_Ch04.indd

95

MTC

02/24/2009 04:07AM

96

Chapter 4

TABLE 4.3

CLASSIFICATION OF STRESSES FOR SOME TYPICAL CASES (ASME, VIII-2)

Vessel

component

Any shell

including cylinders, cones,

spheres, and

formed heads

Location

Shell plate

remote from

discontinuities

Near nozzle or

other opening

Any location

Cylindrical or

conical shell

Dished head or

conical head

Flat head

Perforated

head or shell

Shell distortions

such as outof-roundness and

dents

Any section across

entire vessel

Junction with

head or ange

Crown

Axial thermal

gradient

Net-section axial

force and/or bending moment applied

to the nozzle, and/or

internal pressure

Temperature difference between shell

and head

Internal pressure

Net-section axial

force, bending moment applied to the

cylinder or cone,

and/or internal

pressure

Internal pressure

Internal pressure

Knuckle or

junction to shell

Center region

Internal pressure

Junction to shell

Internal pressure

Typical ligament in

a uniform pattern

Pressure

Isolated or

atypical ligament

ASM_Jawad_Ch04.indd

Origin of stress

Internal pressure

96

Internal pressure

Pressure

Type of stress

Classication

General membrane

Gradient through plate

thickness

Membrane

Bending

Local membrane

Bending

Peak (llet or corner)

P m

Q

Membrane

Bending

Q

Q

Membrane

Bending

Pm

Q

Membrane stress

averaged through the

thickness, remote

from discontinuities;

stress component

perpendicular to

cross-section

Bending stress

through the thickness;

stress component

perpendicular to

cross-section

Membrane

Bending

Membrane

Bending

Membrane

Bending

Membrane

Bending

Membrane

Bending

Membrane (averaged

through cross-section)

Bending (averaged

through width of

ligament, but gradient

through plate)

Peak

Membrane

Bending

Peak

Pm

MTC

Q

Q

PL

Q

F

Pb

PL

Q

Pm

Pb

PL1

Q

Pm

Pb

PL

Q2

P m

Pb

F

Q

F

F

02/24/2009 04:07AM

Vessel

component

Nozzle

(see Section

5.6) of VIII-2

Location

Within the limits

of reinforcement

given by Section

4.5 of VIII-2

of reinforcement

given by Section

4.5 of VIII-2

Nozzle wall

Origin of stress

Type of stress

Pressure and

external loads and

moments including

those attributable to

restrained free end

displacements of

attached piping

Pressure and external axial, shear,

and torsional loads

including those

attributable to

restrained free end

displacements of

attached piping

Pressure and

external loads and

moments, excluding

those attributable to

restrained free end

displacements of

attached piping

Pressure and all

external loads and

moments

Gross structural

discontinuities

General membrane

Bending (other than

gross structural

discontinuity stresses)

averaged through

nozzle thickness

P m

P m

General membrane

P m

Membrane

Bending

PL

Pb

Membrane

Bending

Peak

Membrane

Bending

Peak

Membrane

Bending

Peak

Membrane

Bending

Equivalent linear

stress4

Nonlinear portion of

stress distribution

Stress concentration

(notch effect)

PL

Q

F

PL

Q

F

Q

Q

F

F

F

Q

Differential

expansion

Cladding

Any

Any

Any

Any

Differential

expansion

Radial temperature

distribution3

Any

Any

Classication

F

F

Consideration shall be given to the possibility of wrinkling and excessive deformation in vessels with large diameter/

thickness ratio.

2

If the bending moment at the edge is required to maintain the bending stress in the center region within acceptable

limits, the edge bending is classied as Pb; otherwise, it is classied as Q.

3

Consider the possibility of thermal stress ratchet.

4

Equivalent linear stress is dened as the linear stress distribution that has the same net bending moment as the actual

stress distribution.

Pm

P

J 1

PR i t

(4.11)

The secondary bending stress component, Q , is given by coordinates EG and FH. The value of Q

is determined by integrating area BJF times its moment arm around point J plus area EAJ times its

moment arm around J and then multiplying the total sum by the quantity 6/t 2. This results in a Q

value of

ASM_Jawad_Ch04.indd

97

MTC

02/24/2009 04:07AM

98

Chapter 4

FIG. 4.5

SEPARATION OF STRESSES IN A CYLINDRICAL SHELL

Q =

g

1

6P

(ln g )

( - 1) 2 2 g 2 - 1

(4.12)

The peak stress component, F , is obtained by subtracting the quantities Pm plus Q from the total

stress at a given location. Hence,

Fmax = P

g2+ 1

g2 - 1

- Pm - Q

(4.13)

Example 4.3

A steam drum has an inside diameter of 30 in. and outside diameter of 46 in. The applied pressure

is 3300 psi. Determine the following:

(a)

(b)

(c)

(d)

Membrane stress Pm

Secondary stress Q

Peak stress F

Solution

(a) From Lames Eq. (4.9) with = 23/15 = 1.533, the maximum and minimum stress at the

inside and outside surfaces are

s i = 3300

1.5332 + 1

1.5332 - 1

= 8190 psi

Vo

2330015332 1

4890 psi

ASM_Jawad_Ch04.indd

98

MTC

02/24/2009 04:07AM

Pm =

3300

1.533 - 1

= 6190 psi

633001533

1533 1 2

1533

1

ln1533

2

15332 1

1590 psi

F = s inside - (Pm + Q )

= 8190 - (6180 + 1590)

= 410 psi

4.3.5

Thermal Stress

Most vessels operating at elevated temperatures are subjected to thermal gradients and stress. If

there are rapid thermal transients in thick-walled vessels, the thermal stress may be non-linear through

the thickness of the component and thus require separation into membrane, secondary and peak

stresses. In Table 4.3, the thermal stress in vessel components is generally classied as either secondary or peak depending on whether the thermal stress is general or local as shown in Fig. 4.6. There are

some conditions where the membrane stress due to thermal conditions is classied as primary stress

in the creep range. This condition will be discussed in later chapters.

Generally, thermal stress affects the cycle life of a component and the analysis is discussed in Chapter 5. However, some thermal stress is also evaluated at steady-state conditions at such areas as skirt

and lug attachments.

OPERATING CONDITIONS

ASME Sections I and VIII provide design rules for common components such as shells, heads,

nozzles, and covers. These rules are intended to keep the design stresses Pm, PL, and Pb in the components within allowable stress limits. The design pressure and temperature are assumed as slightly

higher than the operational pressure and temperature and are taken at a point in the operational cycle

where they are at maximum. Many pressurized equipment such as power boilers and hydrocrackers

in reneries operate at essentially a steady-state condition. However, the combination of mechanical

and thermal loading may necessitate additional creep analysis. Such analysis is discussed in Chapters

5 and 6.

Section VIII-2 has specic procedure for setting limits on various combinations of the stresses Pm,

PL, Pb, and Q . The maximum values of these combinations are called equivalent stress values and are

designated by Pm, PL, Pb, and Q as shown in Fig. 4.7. The procedure for calculating equivalent stress

values from stress values is given by the following steps.

1. At a given time during the operation (such as steady state condition), separate the calculated

stress at a given point in a vessel into (Pm), (PL + Pb), and (PL + Pb + Q). The quantity PL

is either general or local primary membrane stress.

ASM_Jawad_Ch04.indd

99

MTC

02/24/2009 04:07AM

100

Chapter 4

Thermal Stress

distribution of temperature or by differing thermal

coefficients of expansion. Thermal stress id

developed in a solid body whenever a volume of

material is prevented from assuming the size and

shape that it normally should under a change in

temperature.

General, Q

Local, F

is associated with almost

complete suppression of

the differential expansion

and thus produces no

significant distortion. Local

thermal stress is classified

as peak stress. Examples of

local thermal stresses are:

1)stress in a small hot spot

in a vessel wall.

2) the difference between

the actual stress and the

equivalent linear stress.

3) the thermal stress in a

cladding material.

classified as secondary stress.

Examples of general thermal

stress are:

1) stress produced by an axial

temperature distribution in a

cylindrical shell.

2) stress produced by the

temperature difference

between a nozzle and the shell

to which it is attached.

3) the equivalent linear stress

produced by the radial

temperature distribution in a

cylindrical shell.

FIG. 4.6

THERMAL STRESS CATEGORIES

2. Each of these three bracketed quantities is actually six quantities acting as normal and

shear stresses on an innitesimal cube at the point selected.

3. Repeat steps 1 through 2 at a different time frame such as startup.

4. Find the algebraic sum of the stresses from steps 1 through 3. This sum is the stress

range of the quantities (Pm), (PL + Pb), and (PL + Pb + Q).

ASM_Jawad_Ch04.indd

100

MTC

02/24/2009 04:07AM

Stress

Category

Description (For

examples,

see Table

5.2 of

VIII-2)

Symbol

Primary

General

Membrane

Local

Membrane

Average stress

across only solid

section.

Considers discontinuities but

not concentrations. Produced

only by mechanical loads.

Average primary

stress across

solid section.

Excludes discontinuities and

concentrations.

Produced only

by mechanical

loads.

Pm

Pm

Secondary

Membrane

plus Bending

Bending

Component of

primary stress

proportional to

distance from

centroid of solid

section.

Excludes discontinuities and

concentrations.

Produced only

by mechanical

loads.

Self-equilibrating

stress necessary

to satisfy continuity of structure.

Occurs at structural discontinuities. Can be

caused by

mechanical load

or by differential

thermal

expansion.

Excludes local

stress

concentrations.

Pb

PL

Peak

1. Increment

added to

primary or

secondary

stress by a

concentration

(notch).

2. Certain thermal stresses

which may

cause fatigue

but not distortion of vessel

shape.

PL + Pb + Q

PL

SPS

1.5S

Use operating loads

PL + Pb

1.5S

PL + Pb + Q + F

Sa

FIG. 4.7

STRESS CATEGORIES AND LIMITS OF EQUIVALENT STRESS (ASME, VIII-2)

5. From step 4, determine the three principal stresses S1, S2, and S3 for each of the stress categories (Pm), (PL + Pb), and (PL + Pb + Q).

6. From step 5, determine the equivalent stress, Se, for each of the stress categories (Pm), (PL +

Pb), and (PL + Pb + Q) in accordance with the strain energy (von Mises) equation

Se =

2

2

2

1

(S1 - S2 ) + (S2 - S3) + (S3 - S1)

(2)1/ 2

1/ 2

(4.14)

The maximum absolute values obtained in step 6 are called equivalent stresses (Pm), (PL +

Pb), and (PL + Pb + Q). The von Mises equation (Eq. 4.14) is for ductile materials and is not

applicable to brittle materials or materials that become brittle during operation. It should also

be noted that for shells subjected to internal pressure the equivalent stress values obtained

from the von Mises theory are about 10% smaller than those obtained from the shear theory

used in Sections III-NB, III-NH, and VIII-2 before 2007.

7. The equivalent stress values in step 6 are analyzed in accordance with Fig. 4.7 for temperatures below the creep range and Fig. 4.8 for temperatures in the creep range. This chapter

ASM_Jawad_Ch04.indd

101

MTC

02/24/2009 04:07AM

102

Chapter 4

FIG. 4.8

FLOW DIAGRAM FOR ELEVATED TEMPERATURE ANALYSIS (ASME, III-NH)

covers load-controlled stress limits as shown on the left-hand side of Fig. 4.8. Strain and

deformation limits as shown on the right side of Fig. 4.8 are covered in Chapter 5 and creep

fatigue is covered in Chapter 6.

The design equivalent stress limits for Pm, PL, and Pb at temperatures below the creep range are

shown by the solid lines in Fig. 4.7. The design equivalent stress is generally limited to

Pm < S for general membrane stress

PL < 1.5S for local membrane stress

(PL + Pb) < 1.5S for local membrane and bending primary stresses

ASM_Jawad_Ch04.indd

102

MTC

02/24/2009 04:07AM

Many of the construction details given in I and VIII for major components meet these design stress

limits, and in general there is no need to run a stress check due to design conditions. The situation in

the creep range is different in that the 1.5S limit on PL and Pb does not apply when loading conditions

are of sufcient duration for creep effects to redistribute elastically calculated stress and strain. This is

reected by the use of the Kt factor in the stress evaluation in III-NH. And although many construction details given in I and VIII for major components meet the design stress limits in the creep range

as given by the rst row in Fig. 4.8, other components such as nozzle reinforcement, lugs, and attachments may require an additional analysis in the creep range in accordance with Fig. 4.8 for steady load

conditions when the design specications require it or when the designer deems it necessary.

For the operating condition, it is necessary to determine the equivalent stress values of Pm, PL,

Pb, and Q. The ASME requirements for dening equivalent stress limits differ substantially for temperatures below the creep zone from those in the creep zone. The criteria below the creep zone for

VIII-2 are given by the dotted lines in Fig. 4.7 and show that primary plus secondary equivalent stress

values are limited by the quantity SPS. This quantity is essentially limited by the larger of 3S or 2Sy,

where S and Sy are the average value for the highest and lowest temperature of the cycle. The VIII-2

methodology for evaluation of primary plus secondary stresses is shown here for convenience as the

VIII-2 methodology does not consider the effects of stress redistribution due to creep. The criteria

for stress in the creep range in accordance with III-NH are shown in the second row of Fig. 4.8. It

involves checking load-controlled then strain-controlled stresses. A detailed description of the III-NH

methodology is given in this chapter for load-controlled stress and in Chapter 5 for strain-controlled

stress.

The above procedure for the separation of stresses is illustrated in the following example.

Example 4.4

Determine the stresses on the inside surface of a thick cylindrical shell subjected to internal pressure.

Solution

Step 1

(a) Circumferential stress

From Eq. (4.11),

P

g - 1

PmH =

(1)

Due to internal pressure away from discontinuities, the quantity PmH = PLH in this case and thus,

PLH =

g - 1

, PbH = 0

PLH + PbH =

P

g - 1

(2)

QH =

6P g

( g - 1) 2

g

1

- 2

(ln g )

g - 1

2

and

PLH + PbH + Q H =

P

( g - 1)

(4 g - 1) -

6g 2

( g 2 - 1)

(ln g )

(3)

From Jawad and Farr (1989),

ASM_Jawad_Ch04.indd

103

MTC

02/24/2009 04:07AM

104

Chapter 4

P

J2 1

PLLc

(4)

PbLc

Q Lc

0

Hence,

P

J2 1

PLLc PbLc

(5)

and

P

J2 1

PLLc PbLc Q c

(6)

PLRc

PbRc

0

P2

Q Rc

(7)

P2

Hence,

PLRc PbRc

P2

(8)

and

PLRc PbRc Q Rc

P

(9)

Step 2

All shearing stress is zero in a cylindrical cylinder subjected to internal pressure.

Step 3

It will be assumed that the initial stresses before application of pressure are all zero.

Step 4

Because the stresses in step 3 are zero, the total sum of the equivalent stress values is given in

step 1.

Step 5

(Pm), which is the same as PL in this case, are given by Eqs. (1), (4), and (7)

S1

P

S

J 1 2

P

S3

J2 1

(10)

P2

The principal stresses for the stress value (PL + Pb) are given by Eqs. (2), (5), and (8)

S1

P

S2

J 1

P

S3

J2 1

P2

(11)

The principal stresses for the stress value (PL + Pb + Q) are given by Eqs. (3), (6), and (9)

S1

S2

ASM_Jawad_Ch04.indd

104

P

6J 2

4 J 1 2

ln J

J 1

J 1 2

P

J2

1

S3

P

(12)

MTC

02/24/2009 04:07AM

Step 6

The stress differences for Pm are obtained from Eq. (10) as

P

J2 1

S1 S2

P J 2 1

S2 S3

2 J 2 1

P J 1

2 J 1

S3 S1

(13)

The stress differences for (PL + Pb) are obtained from Eq. (11) as

S1 S2

PJ

J2 1

P J 2 1

S2 S3

2 J 2 1

PJ 1

2J 1

S3 S1

(14)

The stress differences for (PL + Pb + Q) are obtained from Eq. (12) as

S1 S2

2P J

J 1 2

2 J 1

3J

2

ln J

J 1

J 1

S2 S3

S3 S1

P

J 1 2

PJ 2

J2 1

J 2 2J

6J 2

J2 1

ln J

(15)

Equations (13), (14), and (15) are substituted into Eq. (4.14) to obtain the equivalent stress values

of Pm, (PL + Pb), and (PL + Pb + Q).

Step 7

The equivalent stress values Pm, (PL + Pb), and (PL + Pb + Q) obtained in step 6 are analyzed

in accordance with Fig. 4.7 for temperatures below the creep range and Fig. 4.8 for temperatures in the creep range as explained in this section.

THE CREEP RANGE

Pressure vessel and boiler components operating in the creep range may require, in some instances,

a special analysis in accordance with load-controlled as well as strain-controlled limits. These instances include sudden upset or regeneration conditions, a few operating cycles with temperature

spikes, and design conditions based on life expectancy greater than the assumed 100,000 hours. Moreover, some construction components such as nozzles, support and jacket attachments, and tubes may

require such analysis in equipment operating under steady-state conditions. The procedure for such

analysis is given in this section.

In the creep range, the criteria for equivalent stress limits shown in Fig. 4.8 are based on limiting

the stresses for the design loads and then for the operating loads. In non-nuclear applications, this cor-

ASM_Jawad_Ch04.indd

105

MTC

02/24/2009 04:07AM

106

Chapter 4

responds to the rst two rows in Fig. 4.8, which are the Design Limits and Levels A and B service

limits. The other two limits in the gure, namely, C and D services, are normally not pertinent to

boiler and pressure vessel operations, although they can be specied by the user if so desired.

The Design Limits in the rst row of Fig. 4.8 can be considered satised for head and shell components when the designer uses Section I and VIII details of construction. The reason is that the requirements of Pm < So and PL + Pb <1.5So are fullled when using the details of Sections I and VIII. Other

details such as nozzle reinforcement and jacket attachments are based, in part, on design by rules and

may require an analysis at the discretion of the designer.

In the creep range, the III-NH operating limits in the second row of Fig. 4.8 must be complied with

in a similar fashion as the VIII-2 limits shown in Fig. 4.7 designated by the dotted lines for temperatures below the creep range. The limits are based on (1) load-controlled stresses and (2) strain- and

deformation-controlled stresses. The load-controlled stresses are discussed in this chapter and the

strain-controlled stresses are discussed in Chapter 5.

The load-controlled limits are given by

Pm Smt

(4.15)

PL Pb KSm

(4.16)

>PL Pb Kt@ St

(4.17)

and

For welded construction, the values of Smt and St listed above are further dened as

Smt = lower value of Smt or 0.8(Rw)(Sr)

St = lower value of St or 0.8(Rw)(Sr)

TABLE 4.4

SHAPE FACTORS

I/c1

Z2

Shape factor, K

Bt2/6

Bt2/4

1.5

R2t

4 R2t

1.27

R3/4

4 R3/3

1.7

(16/3)(1 13)/(1 14 )

Ranges between 1.27 and 1.70

Shape

= M/Z for plastic analysis.

1 = Ri/Ro.

1

2

ASM_Jawad_Ch04.indd

106

MTC

02/24/2009 04:07AM

Example 4.5

The vessel shown in Fig. 4.9a is constructed of SA-387-22 Cl.1. The shell is assumed to be thin such

that the stress distribution across the thickness is uniform due to applied pressure. The design and

operating conditions are

Design pressure = 0 psi to 500 psi

Operating pressure = 0 psi to 450 psi

Operating temperature = 70F to 810F

= 0.30

Use 100,000 hours service life.

Ri = 16.00 in.

joint efciency = 1.0

The designer of a VIII-1 vessel is required, occasionally, to check the stresses at various components subsequent to the design. However, VIII-1 does not provide rules for such detailed stress

analysis. Accordingly, the designer uses the rules of VIII-2, part 5, when the temperature is below the

creep range, and the rules of III-NH when the temperature is in the creep range. VIII-2 and III-NH

have substantial differences in how they dene the creep range. This shows up in this example, which

is below the creep range for VIII-2 and in the creep range for III-NH. This procedure is demonstrated

in (b) and (c) below.

(b) Evaluate the operating stress values at the inside surface of the head-to-shell junction using

the criteria of VIII-2, part 5, as a general guide.

(c) Evaluate the operating stress values at the inside surface of the head-to-shell junction in accordance with the criteria for load-controlled stress in NH, as a general guide.

In item (c) above, only the load-controlled stresses will be checked. The strain and deformationcontrolled stresses, which must also be simultaneously evaluated in accordance with Section NH, will

be discussed in Chapter 5.

Solution

(a) Design Condition

From Section II-D, the allowable stress, S, at 825F is 16,600 psi. This temperature is below the

900F set by Section II-D for this material, where creep and rupture criteria control.

FIG. 4.9

HEAD-TO-SHELL JUNCTION

ASM_Jawad_Ch04.indd

107

MTC

02/24/2009 04:07AM

108

Chapter 4

t = PR i /(SEo - 0.6P)

= (500)(16)/[16,000 (1.0) - 0.6(500)]

= 0.49 in.

Use t = in.

The required head thickness is obtained from the following equations.

m

04905

098

C

044m

043

th

D i CP S 12

12

364 in

Details of shell to head junction in accordance with VIII-1 are shown in Fig. 4.9b.

(b) Evaluation of Operating Stresses in Accordance with VIII-2, Part 5

Operating pressure = 450 psi

Operating temperature = 810F

The 810F operating temperature is below the 850F limit set by Section VIII-2, where the allowable stress values are controlled by creep for this material. Hence, creep is not a consideration in accordance with the rules of VIII-2.

From Section II-D for VIII-2 values, SAv = 18,790 psi. (S = 20,000 psi at room temperature and S

= 17,580 psi at 810).

From Table Y-1 of Section II-D, av. Sy = 28,260 psi. (Sy = 30,000 psi at room temperature and Sy

= 26,520 psi at 810F).

From Table TM-1 of Section II-D, E = 26,200,000 psi at 810F.

The unknown moment, Mo, and shear force, Vo, at the junction are shown in Fig. 4.9c. Their values

are obtained from the following two compatibility equations

Deection of shell due to pressure Mo Vo

(1)

and

Rotation of shell due to Mo Vo

(2)

Deection of shell

Due to pressure = 0.85PRm2/Et

Due to moment = Mo/2D2

ASM_Jawad_Ch04.indd

108

MTC

02/24/2009 04:07AM

Rotation of shell

Due to pressure = 0

Due to moment = -Mo/D

Due to radial load = Vo/2D2

Radial deection of at head

Due to pressure = 0

Due to moment = 0

Due to radial load = RmVo/Eth

Rotation of at head

Due to pressure = -3(1 - )PRm3/2Eth3

Due to moment = 12(1 - )RmMo/Eth3

Due to radial load = 0

where

D = Et3/12(1 - 2)

E = modulus of elasticity

Rm = mean radius

t = thickness of shell

th = thickness of at head

= [3(1 - 2)/Rm2t2]0.25

= Poissons ratio

Solving the two compatibility equations (Eqs. 1 and 2) for Mo and Vo gives

Mo = 1283 in.-lb/in.

and

Vo = 994 lbs./in.

It is assumed that the initial stresses at the beginning of the operating cycle are zero. Hence, the

following calculations can be treated as stress range values rather than stress values. Also, only the

equivalent stress quantity (PL + Pb + Q) is required to be checked for the operating conditions in accordance with Fig. 4.7.

Shell stress range calculations

Axial stress

P L = PR i /2t = 7200 psi

(a)

Pb = 0

(b)

(c)

(d)

PL = 0 psi

(e)

Pb = 0

(f )

(g)

Circumferential stress

ASM_Jawad_Ch04.indd

109

MTC

02/24/2009 04:07AM

110

Chapter 4

PLc Pbc Qc

9240 psi

(h)

Radial stress

PLc 225 psi

Pbc

(i)

(j)

Qc 225 psi

(k)

(l)

The three principal stresses for (PL + Pb + Q) at the inside surface of the shell at the head-to-shell

junction are then given by Eqs. (d), (h), and (l).

S1 38000 psi

S2

9240 psi

S3 450 psi

Thus, the maximum equivalent stress value of (PL + Pb + Q) is obtained from Eq. (4.14) as

0707 28760 2 9690 2 38450 2

PL Pb Q

1 2

34600 psi

2Sy

228260

3S

318790

56520 psi

or

56370 psi

Thus, the calculated value of 34,600 psi is less than the allowable stress of 56,520 psi.

Thus, based on an elastic analysis below the VIII-2 creep range, the stress at the inside surface of

the shell at the shell-to-head junction is adequate.

The above calculations satisfy the stress requirements of VIII-2.

The following calculation for secondary stress, Q, is detailed here but is not required in the VIII-2 or

the III-NH load-controlled Stress evaluations. It is performed here for expediency for use in Chapter

5 in the strain and deformation limit calculations.

(Q)

From Eqs. (c), (g), (k), and (4.14),

Maximum equivalent stress value of Q

27540 psi

Radial stress

PLc

Pbc

Vth

265 psi

6Mth2

Qc

550 psi

0 psi

PLc Pbc Qc

ASM_Jawad_Ch04.indd

110

(m)

(n)

(o)

815 psi

MTC

(p)

02/24/2009 04:07AM

Tangential stress

PLc

Pbc

0 psi

03550

Qc

(q)

165 psi

(r)

0 psi

PLc Pbc Qc

(s)

165 psi

(t)

PLc 225 psi

Pbc

(u)

(v)

Q 225 psi

(w)

(x)

The three principal stresses at the inside surface of the shell at the head-to-shell junction are then

given by Eqs. (p), (t), and (x).

S1

815 psi

S2

165 psi

S3 450 psi

Thus, the maximum equivalent stress value of (PL + Pb + Q) is obtained from Eq. (4.14) as

(PL + Pb + Q) = 0.707 (815 - 165 )2 + [165 - (-450 )] 2 + (-450 - 815 ) 2

1/ 2

= 1095 psi

By inspection, these equivalent stress values are well below the allowable stress limits.

The stress at the middle of the head is equal to the stress in the middle assuming a simply supported

head minus the stress caused by the edge moment due to the shell restraint.

Pb

33 P PR 2m 8th2 550

10460 550

9910 psi 15Sm

The edge stress can be classied as either primary bending stress, Pb, or secondary stress, Q,

depending on the design of the head. If the edge moment is used to reduce the moment in the middle

of the head in order to meet the allowable stress criterion, 1.5S, then it must be classied as Pb. If the

stress in the middle of the head is based on a simply supported head without taking the edge stress into

consideration, then the edge stress is classied as secondary.

The following calculation for secondary stress, Q, is detailed here but is not required in the VIII-2 or

the III-NH load-controlled stress evaluations. It is performed here for expediency for use in Chapter

5 in the strain and deformation limit calculations.

From Eqs. (o), (s), (w), and (4.14),

Maximum equivalent stress value of Q

225 psi

ASM_Jawad_Ch04.indd

111

MTC

02/24/2009 04:07AM

112

Chapter 4

From Table I-14.3D of III-NH, Smt = 17,100 psi at 810F for 100,000 hours.

From Table I-14.4D of III-NH, St = 17,200 psi at 810F for 100,000 hours.

From Table I-14.10, R = 1.0 at 810F.

From Table I-14.6D, Sr = 25,800 psi at 810F for 100,000 hours.

Smt

17,100 psi, or

0.8(1.0)(25,800)

Smt

20640 psi

17100 psi

St

It should be noted that in part (b) of this example, only the equivalent stress (PL + Pb + Q) is required to be checked for the operating conditions in accordance with Fig. 4.7. However, in the

creep range under the rules of NH, the quantities Pm, (PL + Pb), (PL + Pb/Kt), Q, and (PL + Pb + Q)

need to be checked in accordance with the requirements of Fig. 4.8. In this chapter, only the quantities Pm, (PL + Pb), and (PL + Pb/Kt) will be evaluated. The remaining quantities, Q, and (PL + Pb +

Q) will be calculated in this example but discussed in Chapter 5.

Shell stress calculations

In the shell and head calculations, values of Pm and PL will be conservatively assumed as similar.

These values are obtained from the von Mises equation given in VIII-2. In addition, these values are

used in the following NH calculations with the understanding that NH uses the Tresca criteria. Such

a hybrid approach enables the designer to use the NH concepts while complying with the VIII-2 philosophy of calculating equivalent stresses.

(Pm)

From Eqs. (a), (e), (i), and (4.14),

Maximum equivalent stress value of Pm

7315 psi

(PL + Pb)

From Eqs. (a + b), (e + f), (i + j), and (4.14),

Maximum equivalent stress value of PL Pb

7315 psi

(PL + Pb/Kt)

From Table 4.4, K = 1.5

Kt

1 K 2

125

Maximum equivalent stress value of (PL + Pb/Kt) = 7315 psi

From Eq. (4.15)

Pm Smt

7315 17,100

ASM_Jawad_Ch04.indd

112

ok

MTC

02/24/2009 04:07AM

PL P b KS m

7315 1516600 ok

PL Pb Kt S t

7315 17,200

ok

In the shell and head calculations, values of Pm and PL will be conservatively assumed as similar.

(Pm)

From Eqs. (m), (q), (u), and (4.14)

Maximum equivalent stress value of Pm = 425 psi

(PL + Pb)

From Eqs. (m + n), (q + r), (u + v), and (4.14)

Maximum equivalent stress value of (PL + Pb) = 910 psi

(PL + Pb/Kt)

From Table 4.4, K = 1.5

Kt

1 K 2

125

Maximum equivalent stress value of (PL + Pb/Kt) = 810 psi

4.6

The separation of stresses, at a given point in a pressure vessel part, into primary and secondary

components is a very tedious task, as illustrated in Section 4.4. Whenever possible, many designers

rely on other approximate approaches to design components. One such approach is the reference

stress method. The reference stress method is a convenient means to characterize the stress in a component operating in the creep range without having to separate the stresses into primary and secondary components. The method is based on limit analysis, as an asymptotic solution, to obtain a stress

in a component.

The rationale for using limit analysis in determining stress in pressure components operating in the

creep range was discussed in Chapters 1 and 2. The relationship between strain rate and stress in the

creep range is normally taken as

de /dT = k s n

(4.18)

where

d/dT = strain rate

k

= constant

n

= creep exponent, which is a function of material property and temperature

= stress

ASM_Jawad_Ch04.indd

113

MTC

02/24/2009 04:07AM

114

Chapter 4

It was shown in Chapter 1 that a stationary stress condition is reached after a few hours of operating a component in the creep range under a constant strain rate. The relationship between strain and

stress can then be taken as

e = K sn

(4.19)

where

= strain

K = constant

4.6.1

Cylindrical Shells

The governing equation for the circumferential stress in a cylindrical shell due to internal pressure in the

creep range under a stationary stress condition (Finnie and Heller, 1959) is derived from Eq. (4.19) as

S = P

2- n

n

(Ro /R )

2/n

+1

g 2/n - 1

(4.20)

where

P

R

Ri

Ro

S

= internal pressure

= radius at any point in the shell wall

= inside radius

= outside radius

= circumferential stress

= Ro/Ri

A plot of this equation for n > 2 is shown by line EF in Fig. 4.10. This line shows that when the

component is operating in the creep range, the stress at the inside surface decreases and the stress at

the outside surface increases compared to the stress in the non-creep region as illustrated by line AB

from Eq. (4.8).

FIG. 4.10

ELASTIC VS. STATIONARY CREEP STRESS DISTRIBUTIONS

ASM_Jawad_Ch04.indd

114

MTC

02/24/2009 04:07AM

A comparison of Eqs. (4.8) and (4.20) is given in Table 4.5. For n = 1, the creep (Eq. 4.20) and the

elastic (Eq. 4.8) equations yield the same answer as they should. For n values between 1 and 2, the

maximum stress given by Eq. (4.20) is on the inside surface of the shell similar to the elastic equation.

However, for n values greater than 2 the maximum stress given by Eq. (4.20) is on the outside surface

of the shell. Hence, creep fatigue could occur on either surface depending on the material value of n.

However, for most materials used in pressure vessel construction, the n value is greater than 2 and

the maximum stress in the creep regime is on the outside surface as shown in Table 4.5. The designer

should be aware of the stress reversal from the inside to the outside surface when evaluating creep and

fatigue stress. Table 4.5 also shows that the coefcients obtained from Eq. (4.20) reach an asymptotic

value for large values of n. This fact is crucial when the designer decides to substitute inelastic analysis

for creep analysis.

The following example illustrates the application of Eq. (4.20).

Example 4.6

A cylindrical shell is constructed from annealed 2.25Cr-1Mo steel and has an inside diameter of 48

in. The internal pressure is 4000 psi and the design temperature is 1000F. Let Eo = 1.0. The isochronous curves for this material are shown in Fig. 4.11.

(a) Determine the required thickness using Section VIII-1 criterion.

(b) Determine the required thickness using Eq. (4.20) and 100,000-hour life.

Solution

(a) Division VIII-1

The allowable stress from Section II-D is 7800 psi. Since P > 0.385S, thick shell equations must be

used.

Z

SEo P

SEo P

7800 40007800 4000

311

TABLE 4.5

STRESS COEFFICIENTS, S/P, FOR SHELL EQUATIONS AND THICKNESSES

Ro/Ri

1.67

1.25

1.1

Shell type

Thick

Intermediate

Thin

Location

Inside Outside

Inside

Outside

Inside

Outside

n

Lames equation

(Eq. 4.8)

Creep equation

(Eq. 4.20)

2.12

1.12

4.56

3.56

10.52

9.52

2.12

1.12

4.56

3.56

10.52

9.52

1.5

2

5

10

50

100

1.69

1.49

1.16

1.05

0.97

0.96

1.36

1.49

1.76

1.85

1.93

1.94

1.95

4.18

4.00

3.68

3.58

3.50

3.49

3.85

4.00

4.28

4.38

4.46

4.47

4.48

10.17

10.00

9.69

9.59

9.51

9.50

9.84

10.00

10.29

10.39

10.47

10.48

10.49

Inelastic equation

(Eq. 4.22)

ASM_Jawad_Ch04.indd

115

MTC

02/24/2009 04:07AM

116

Chapter 4

FIG. 4.11

AVERAGE ISOCHRONOUS STRESS-STRAIN CURVES (ASME, III-NH)

Ri Z1 2 1

24(1.7635 1)

1832 in.

It should be noted that the allowable stress of 7800 psi is based, in part, on rupture at 100,000

hours.

ASM_Jawad_Ch04.indd

116

MTC

02/24/2009 04:07AM

A value of n = 6.26 is obtained from points A and B in Fig. 4.11 and Eq. (4.19).

From Eq. (4.20) with R = Ro,

7800

4000

195

2 nn Ro R 2 n 1

RoRi 2n 1

0680510 2 626 10

Ro240

2626

10

or

Ro

t

3856 in.

Ro R

1456 in.

The thickness, which is based on stationary creep consideration, is about 25% less than the thickness obtained from VIII-1. This is because VIII-1 uses Lames Eq. (4.9), which gives a high stress

value at the inner surface, whereas Eq. (4.20) redistributes the stresses resulting in a maximum stress

at the outside surface that is lower than Lames high stress at the inner surface as shown in Fig. 4.10.

The last entry in Table 4.5 is based on the maximum pressure allowed in a cylinder based on limit

analysis. The stress on the inside surface of a cylinder due to pressure when the cylinder begins to yield

on the inside surface (Chen and Zhang, 1991) is given by

S

P J 2 1 J 2 1

(4.21)

whereas the maximum stress on the outside of a cylinder when the wall is completely yielded is given

by

S

P ln J

(4.22)

where

P*

P**

S

= pressure when cylinder wall completely yields

= stress

= Ro/Ri

The last entry in Table 4.5 is based on the limit analysis Eq. (4.22). Results are almost identical

to those obtained from Eq. (4.20) with large values of n. This conclusion is used to justify the use of

limit analysis for creep evaluations and is the basis for the Reference Stress method used in industry

(Larsson, 1992). The general equation for the reference stress is given by

SR

PPu S y

(4.23)

where

P

Pu

SR

Sy

= applied load

= ultimate load assuming rigid perfectly plastic material

= reference stress

= yield stress

ASM_Jawad_Ch04.indd

117

MTC

02/24/2009 04:07AM

118

Chapter 4

The procedure for using the stress reference method consists of rst assuming a thickness or conguration of a given part and then performing a limit analysis. A reference stress is then obtained from

the limit analysis and compared to an allowable stress for the component. The procedure is based on

trial-and-error to obtain a conguration that is stressed within the allowable values.

The main method of obtaining reference stress is by using a numerical procedure such as a nite

element analysis. The reason for this is that closed-form solutions are only available for a handful of

components such as shells, at covers, and beams.

It is noteworthy that, at present, there is no general consensus among engineers regarding a stress reference equation for thermal loads (Goodall, 2003) or one that combines pressure and thermal loads.

Example 4.7

What is the required thickness of the heat exchanger channel shown in Fig. 4.12? The material is

304 stainless steel. The design temperature is 1150F and the design pressure is 2500 psi. The thickness is to be calculated in accordance with

(a)

(b)

(c)

(d)

creep Eq. (4.20)

limit Eq. (4.22)

reference stress Eq. (4.23)

Solution

(a) ASME Section VIII-1

The allowable stress, S, from II-D is 7700 psi.

The required thickness from VIII-1 is

t

PR i SEo 06P

2500 u 187700 u 10 06 u 2500

726 in.

18 72618

14033

FIG. 4.12

HEAD EXCHANGER CHANNEL

ASM_Jawad_Ch04.indd

118

MTC

02/24/2009 04:07AM

From Eq. (4.8), the stress at the inside and outside surfaces are

Vi

7660 psi

Vo

2500 2 14033 2 1

5160 psi

Both of these values are below the allowable stress of 7700 psi.

(b) Creep Eq. (4.20)

Equation (4.20) can be solved once the value of n is known. The value of n is also needed in evaluating creep and fatigue analysis. Section VIII does not list any n values. However, an approximate value

of n for a limited number of materials can be obtained from a stress-strain curve at a given temperature

for a given number of hours listed in III-NH. Such curves are presented in isochronous charts. One

such chart is shown in Fig. 4.13. The value of n is determined by using Eq. (4.19) in conjunction with

any of the curves in the chart. Hence, if points A and B are chosen on the hot tensile curve, then the

two unknowns, K and n, in Eq. (4.19) can be calculated. The result is n = 6.1. Other curves in the

chart could be used as well. For example, using points C and D on the 300,000-hour curve results in a

calculated value of n = 5.8. This slight variation in the calculated values of n from various curves has

no practical signicance in calculating stresses as illustrated in Table 4.5 Using an average value of

n = 6.0 in Eq. (4.20) results in the following stress at the inside and outside surfaces:

S inside

5290 psi

Soutside

6940 psi

The results indicate that there is about a 10% relaxation in the maximum stress values when the operating temperature increases from non-creep to the creep regime. This stress reduction is signicant

when evaluating fatigue and creep life.

The required thickness is obtained by using n = 6.0 in Eq. (4.20). This yields

7700

2500

2 6

6

102 6 1

J 26 1

or

J

13611

Ro

2450

65 in.

Accordingly, a thickness of 7.26 in. will result in a lower stress and a longer fatigue life than is implied by VIII-1.

(c) Limit analysis Eq. (4.22)

7700

2500 ln J

or

ASM_Jawad_Ch04.indd

119

13836

Ro

2490

69 in.

MTC

02/24/2009 04:07AM

120

Chapter 4

FIG. 4.13

AVERAGE ISOCHRONOUS STRESS-STRAIN CURVES (ASME, III-NH)

ASM_Jawad_Ch04.indd

120

MTC

02/24/2009 04:07AM

The procedure for the reference stress method is to rst assume a thickness and then calculate the

limit pressure in a component. Equation (4.23) is then used to obtain the reference stress. In this case,

however, we used in part (c) a closed form solution to obtain a thickness of 6.9 in. The limit pressure

Pu based on this trial thickness is calculated from Eq. (4.22)

J

18 6918

Sy ln J

13833

Sy ln 13833

03245 Sy

SR

PPu Sy

250003245 Sy Sy

7700 psi

4.6.2

Spherical Shells

The behavior of spherical shells due to internal pressure is very similar to that of cylinders discussed

in Section 4.6.1. Lames equation for the stress in a thick spherical shell (Den Hartog, 1987) due to

internal pressure is expressed as

S

12 Ro R 3 1

(4.24)

Ro Ri 3 1

The governing equation for the circumferential stress in a spherical shell due to internal pressure

in the creep range under a stationary stress condition (Finnie and Heller, 1959) is derived from Eq.

(4.19) as

3 2n

2n

Ro R

Ro Ri

3n

3n

1

(4.25)

1

The governing equation for the maximum circumferential stress in a spherical shell due to internal

pressure using plastic theory (Hill, 1950) is expressed as

S

1

1

2 ln Ro Ri

(4.26)

Equation (4.25) reverts to Eq. (4.24) when n = 1.0. The value of S in Eq. (4.25) approaches that

given by Eq. (4.26) as n increases to innity.

Example 4.8

A spherical head has an inside radius of 24 in., a thickness of 0.60 in., and is subjected to an internal

pressure of 1000 psi. Calculate the following:

(a) Maximum elastic stress from Eq. (4.24)

(b) Maximum stress from Eq. (4.25) with n = 1, 6, 100

(c) Maximum stress from Eq. (4.26)

ASM_Jawad_Ch04.indd

121

MTC

02/24/2009 04:07AM

122

Chapter 4

Solution

(a) From Eq. (4.24) with R = 24 in.,

S

1000

1224624 3 1

24624 3 1

20000 psi

S

100

S

1000

1

1

2 ln24624

19250 psi

A comparison of the three methods shows that spherical shells behave essentially the same as cylindrical shells.

4.7

The omega method was developed by the Metal Properties Council for the American Petroleum

Institute (API) and published in the API 579 Fitness for Service document. It is intended to calculate

the remaining life, in hours, of a pressure component operating in a renery at a given temperature and

stress level. The background of the method is described by Prager (Prager, 1995) and the analysis procedure is given in API 579 (API, 2000). The procedure is based on a modied Nortons law equation

and supplemented by creep data obtained for various materials used by API. Some of the parameters

needed in the analysis are selected by the designer and are based on judgment and experience. The

value of these parameters can drastically affect the results of the analysis.

The procedure consists of rst calculating the three principal stresses, S1, S2, and S3, at a given location in the vessel. The effective stress based on the strain energy method is then obtained from

Se

(4.27)

The omega uniaxial damage parameter, W, is then calculated from a polynomial obtained from test

data and is expressed as

log10 W = (C o + D cd ) + Te (C l + C 2 S + C 3 S2 + C 4 S3 )

(4.28)

where

C0, C1, C2, C3, C4 = material constants obtained from API 579 for various materials (Table 4.6 gives

a sample of C values for 2.25Cr-1Mo steel in the annealed condition)

SA = log10 (Se)

T = operating temperature (F)

Te = 1/(T + 460)

Dcd = creep ductility factor, which ranges from +0.3 for brittle behavior to -0.3 for ductile behavior.

ASM_Jawad_Ch04.indd

122

MTC

02/24/2009 04:07AM

TABLE 4.6

CONSTANT VALUES FOR

ANNEALED 2.25CR-1MO STEEL

(COURTESY OF API)

Constant

Value

C0

C1

C2

C3

C4

C5

C6

C7

C8

C9

1.85

7,205.0

2,436.0

0.0

0.0

21.86

51,635.0

7,330.0

2,577.0

0.0

The strain rate exponent, n, and the initial strain rate, co, are calculated from the polynomial equations

f = -Te (C 7 + 2C 8 S + 3C 9 S2 )

(4.29)

log10 e co = [(C 5 + D sr ) + Te (C 6 + C 7 S + C 8 S2 + C 9 S3 )]

(4.30)

and

where

C5, C6, C7, C8, C9 = material constants obtained from API 579 for various materials (Table 4.6 gives

several samples of C values for 2.25Cr-1Mo steel in the annealed condition)

Dsr = scatter material factor, which ranges from +0.5 for top of the scatter band to -0.5 for the bottom of the scatter band.

The omega multiaxial damage parameter, Wm, is calculated from the equation

:m

G 1

:n

DI

(4.31)

where

Wn = max[(W - n), 3.0]

= 0.33{[(S1 + S2 + S3)/Se] - 1.0}

= 3.0 for heads

= 2.0 for cylinders and cones

= 1.0 for other components

The remaining life, L, of a component at a given stress level and temperature is then calculated from

the quantity

L

: m H coc 1

(4.32)

The quantities Dcd and Dsr in the above analysis are based on judgmental evaluation of the designer.

Slight variations in these values will change the calculated remaining life, L value, substantially. The

following example illustrates the application of this method to a vessel shell in a renery.

Example 4.9

A pressure vessel has an inside radius of 10 ft and a shell thickness of 1.23 in. Material of construction is 2.25Cr-1Mo annealed steel. The inside pressure is 80 psi and the operating temperature is

ASM_Jawad_Ch04.indd

123

MTC

02/24/2009 04:07AM

124

Chapter 4

1000F. Inspection of the material indicates that a ductility factor, Dcd, of 0.0 is to be used. Moreover,

the value for material scatter band, Dsr, of -0.5 is deemed applicable. Determine the expected life of

the shell.

Solution

The stress values in the shell are

S1

78 ksi

S2

S3

39 ksi

008 ksi

07071 39 2 398 2 7882

Se

1 2

6824 ksi

S = log10 (6.824 ) = 0.834

11000 460

Te

68493 u 10 4

log10 :

16934

and

:

49363

I

log10 H coc

7951

^2186 05 68493 u 10 4 >51635 73300834

25770834 2 0@`

75915

and

H coc

25613 u 10 8

The omega multiaxial damage parameter, Wm, is calculated from Eq. (4.31) as

:n

41412

02319

:m

1141

The remaining life, L, of the vessel is then calculated from Eq. (4.32)

L

1

ASM_Jawad_Ch04.indd

124

MTC

02/24/2009 04:07AM

It is of interest to note that if the material condition Dcd was chosen to be +0.3 then the life expectancy would have reduced to 16.3 years. On the other hand, if Dcd was chosen as -0.3, then the

life expectancy would have increased to 92.4 years. Thus, selecting the correct value of Dcd becomes

crucial in the evaluation of life expectancy in this method. API 579 does not give guidelines regarding

the proper selection of Dcd or Dsr.

Problems

4.2 A pressure vessel is supported on legs. The membrane stresses, Pm, in the shell at the vicinity

of the legs when the vessel is not pressurized are:

Condition 1

c = 0 psi

= -700 psi

c = 350 psi

r = 0 psi

The total membrane stresses in the shell when the vessel is pressurized are:

Condition 2

c = 1800 psi,

= 200 psi,

tc = 350 psi,

r = -150 psi

Comment

The shear stresses cr and 1r are equal to zero in this case. Moreover, the principal stresses in the cl

plane are obtained from the equation

s 1,2 =

( s c + s )

s c - s

2

1/ 2

2

2

+ t c

(A)

4.2 An internal tray has a diameter of 60 in. and is subjected to a 2-psi differential pressure. The

design temperature is 1000F and the material is SA 387-22 cl2.

Data:

The tray is assumed simply supported.

S = 8000 psi at 1000F from VIII-1.

From Section II-D for VIII-2 values, SAv = 14,000 psi (S = 20,000 psi at room temperature and

S = 8000 psi at 1000F).

From Table Y-1 of Section II-D, av. Sy = 26,850 psi (Sy = 30,000 psi at room temperature and

Sy = 23,700 psi at 1000F).

So = 8000 psi at 1000F from III-NH.

Sm = 15,800 psi from Fig. I-14.3D of III-NH.

From Table I-14.3D of III-NH, Smt = 5200 psi at 1000F for 300,000 hours.

From Table I-14.4D of III-NH, St = 5200 psi at 1000F for 300,000 hours.

Perform the following:

(a) Determine the thickness in accordance with VIII-1.

(b) Evaluate the stresses in accordance with VIII-2.

(c) Evaluate the stresses in accordance with Load-Controlled criteria of III-NH based on

300,000 hours.

The elastic bending stress of a simply supported plate is given by

= 1.24PR2/t 2

ASM_Jawad_Ch04.indd

125

MTC

02/24/2009 04:07AM

CREEP-FATIGUE CRACK IN A 2.25 Cr-1Mo STEEL TUBE (COURTESY OF AMEREN UE, ST. LOUIS, MISSOURI)

ASM_Jawad_Ch05.indd

126

MTC

02/24/2009 03:56AM

CHAPTER

5

ANALYSIS OF COMPONENTS:

STRAIN- AND DEFORMATIONCONTROLLED LIMITS

5.1

INTRODUCTION

Stress analysis of a component in the creep regime requires both load- and strain-controlled evaluation. Load-controlled limits were discussed in Chapter 4 and the strain-controlled limits are described

in this chapter. The evaluation for the strain-controlled limits is made in accordance with one of three

criteria detailed in III-NH. The rst criterion, based on elastic analysis, is the easiest to use but is

very conservative. When the stress calculations cannot satisfy the elastic limits then the component

is redesigned or the analysis proceeds to the second criterion based on simplied inelastic analysis.

This second criterion is costlier to perform than the elastic analysis and requires additional material

data. However, the results obtained from the simplied inelastic analysis are more accurate than those

obtained from an elastic analysis.

When the simplied inelastic stress analysis cannot satisfy the simplied inelastic stress limits, then

the component is redesigned or the analysis proceeds to the third criterion based on inelastic analysis.

To perform inelastic analysis, it is necessary to dene the constitutive equations for the behavior of the

material in the creep regime. These equations are not readily available for all materials. As a consequence, this third criterion is costly to perform but gives reasonably accurate results once performed.

5.2

The procedure for strain and deformation limits is intended to prevent ratcheting. Section III-NH

gives the designer the option of using one of three methods of analysis. They are the elastic, simplied inelastic, and inelastic analyses. The object of all of these methods is to limit the strains in the

operating condition to 1% for membrane, 2% for bending, and 5% for local stress. At welds, the allowable strain is one half those values. It is of interest to note that some engineers believe that in some

instances the limitation set on primary membrane strain of 1% may be too conservative because this

criterion does not affect the overall failure of the component. Thus, a 1% strain in a ange may result

in an unwanted leakage, whereas the same strain at the junction of a at head to shell junction may

be acceptable.

The elastic method, which is very conservative, is generally applicable when the primary plus secondary stresses are below the yield strength. The simplied inelastic analysis, which has less conservatism built into it compared to the elastic analysis, is based on bounding the accumulated membrane

strain. The last option is to perform an inelastic analysis. The inelastic analysis yields accurate results

127

ASM_Jawad_Ch05.indd

127

MTC

02/24/2009 03:56AM

128

Chapter 5

but has the drawback of being expensive and time consuming to perform. It requires a large amount of

material property data that may not be readily available for the material under consideration.

One potential disadvantage of using the strain- and deformation-controlled limits of III-NH is that it

requires separate treatment of primary and secondary stress categories as shown in Fig. 4.8. This condition is avoided in VIII-2 by combining the primary and secondary stress categories into one quantity.

For simple structures, the separation of primary and secondary stress in III-NH may not be a big problem. But for more complex structures with asymmetrical geometry and loading, it can be difcult, if not

impossible, to sort out primary and secondary stress categories from a detailed nite element analysis.

Section III-NH, however, does give the designer the option (called test A-3) to use the combination

of primary plus secondary stress limits similar to VIII-2 when the effects of creep are negligible. This

creep modied shakedown limit avoids the potential problem of separating primary and secondary

stresses and is used as an alternative to the standard III-NH strain and deformation limit.

5.3

ELASTIC ANALYSIS

The strain and deformation limits in the elastic stress analysis are considered to be satised if they

meet the requirements of A-1, A-2, or A-3. A summary of the requirements for these three tests is

shown in Fig. 5.1. The membrane and primary bending stresses used in these tests are dened as

X

PL Pb Kt Sy

(5.1)

Y

QSy

(5.2)

The value of Kt in Eq. (5.1) is approximated by III-NH to a value of 1.25 for rectangular crosssections. The actual value of Kt for pipes and tubes is less than 1.25. Table 4.4 shows the Kt values for

rectangular and circular cross-sections.

Important denitions applicable to tests A-1 and A-2

X includes all primary membrane and bending stresses.

Y includes all secondary stresses.

Sy is the average of the yield stress at the maximum and minimum temperatures during the

cycle.

5.3.1

Test A-1

This test applies for cycles where both extremes of the cycle are within the creep range of the material (see test A-2 for denition of creep range for this application). The governing equation is

X Y Sa Sy

(5.3)

(a) 1.25St using the highest wall averaged temperature during the cycle and time value of 104

hours

(b) the average of the two Sy values associated with the maximum and minimum wall averaged

temperatures during the cycle.

The requirement of item (a) is based on the concern that creep relaxation at both ends of the

temperature cycle will exacerbate the potential for ratcheting. Experience has shown that using a

value of St at 104 hours as a stress criterion is a realistic assumption. The requirement of (b) is based

ASM_Jawad_Ch05.indd

128

MTC

02/24/2009 03:56AM

No

Yes

Test A-3

temperature zone (see Table 5-1)

_ lesser of

(Pm + Pb + Q) <

3 Sm or 3 S m

No

Yes

Part of cycle falls below NB to

NH temperature boundary

Test A-1

Test A-2

Yes

No

_ 1.0

(X + Y ) <

_ Smaller of

(X + Y ) <

(1.25 St / Sy) or 1.0

3 S m = 1.5 Sm + SrH

1.5 Sm + St/2

3 S m = SrL + SrH

StL/2 + StH/2

FIG. 5.1

STRAIN-CONTROLLED LIMITS ELASTIC CRITERION

ASM_Jawad_Ch05.indd

129

MTC

02/24/2009 03:56AM

130

Chapter 5

on averaging the yield stress associated with the maximum and minimum temperatures as a good

approximation.

5.3.2

Test A-2

This test is applicable for those cycles in which the average wall temperature at one of the stress extremes dening the maximum secondary stress range, Q, is below the temperature given in Table 5.1.

X Y 10

(5.4)

Values shown in Table 5.1 are the approximate temperatures above which the material allowable

stress values at 100,000 hours are controlled by creep and rupture.

5.3.3

Test A-3

This test, although applicable to all conditions, was originally intended for components that are in

the creep range for only a portion of their expected design life. Compliance with this test indicates that

creep is not an issue and the rules of VIII-2 may be used directly. The calculated stresses are satised

when all of the following criteria are met:

(a) The combined primary and secondary stress are limited to the lesser of

PL Pb Q 3Sm

(5.5)

(If Q is due to thermal transients then Sm is the average of values taken at the hot and cold ends of

the cycle. If pressure-induced loading is part of Q then Sm is the value at the hot end of the cycle.)

or

(5.6)

PL Pb Q 3Sm

where

3Sm = (1.5Sm + SrH). If both temperature extremes of the cycle are in the creep regime, then the lower

temperature relaxation strength, SrL, should be substituted for 1.5Sm. The extremes of the cycle are considered to be in the creep regimen when the temperature is greater than 700F for

2.25Cr-1Mo and 9Cr-1Mo-V steels and greater than 800F for stainless and nickel alloys.

SrH = hot relaxation stress. It is obtained by performing a pure uniaxial relaxation analysis starting with an initial stress of 1.5Sm and holding the initial strain throughout the time interval

equal to the time of service in the creep regime. An acceptable, albeit conservative, alternate to this procedure is to use the elevated temperature creep dependent allowable stress

level, 0.5St, as a substitute for SrH.

St = temperature- and time-dependent stress intensity limit.

(b) Thermal stress ratcheting must be kept below a certain limit dened by the equations

TABLE 5.1

TEMPERATURE LIMITATION

Material

Temperature, F (C)

316 stainless steel

Alloy 800H

2.25Cr-1Mo steel

9Cr-1Mo-V

ASM_Jawad_Ch05.indd

130

948 (509)

1011 (544)

1064 (573)

801 (427)

940 (504)

MTC

02/24/2009 03:56AM

y = 1/x

y

41 x

(5.7)

(5.8)

where

x = (general membrane stress, Pm)/Sy

y = (range of thermal stress calculated elastically, DQ)/Sy

(c) Time duration limits are given by

Ti Tid

01

(5.9)

where

Ti = total time duration during the service lifetime of a component at the highest operating

temperature

Tid = maximum allowable time as determined by entering stress-rupture chart at temperature Ti

and a stress value of 1.5 times the yield stress at temperature Ti

(d) Strain limit is given by

Hi

02

(5.10)

where

i = creep strain that would be expected from a stress level of 1.25Sy|Ti applied for the total duration of time during the service lifetime that the metal is at temperature Ti

_

Note: Table 5.1 in test A-2 has different temperatures than those dening (3S m) in

test A-3 for the onset of creep effects.

Example 5.1

An VIII-1 vessel has a design temperature of 1100F and design pressure of 315 psi. Material of

construction is stainless steel grade 304. The operating temperature is 1050F and the operating pressure is 300 psi. The required shell thickness due to design condition is 1.0 in. However, the thickness

of part of the shell is increased to 2.0 in. to accommodate the reinforcement of some large nozzles in

the shell. The junction between the 1.0-in. and 2.0-in. shells is shown in Fig. 5.2a. Temperatures at the

inside and outside surfaces of the shell during the steady-state operating condition are shown in Fig.

5.2b. The operating cycle (Fig. 5.2c), is as follows:

Pressure

The pressure increases from 0 psi to 300 psi in 24 hours. It remains at 300 psi for about 18 months

(13,000 hours), and is then reduced to zero in 1 hour. The same cycle is then repeated after a shutdown

of 2 weeks for maintenance.

Temperature

The temperature increases from ambient to 1050F in 24 hours. It remains at 1050F for about 18

months (13,000 hours) and then reduced to ambient in 4 days. The same cycle is then repeated after

the shutdown.

Evaluate the following:

1. Stress away from any discontinuities due to design condition.

2. Stress away from any discontinuities due to operating condition using elastic analysis.

ASM_Jawad_Ch05.indd

131

MTC

02/24/2009 03:56AM

132

Chapter 5

FIG. 5.2

VESSEL GEOMETRY AND OPERATING CYCLES

Data:

So = 9800 psi at 1100F (Table I-14.2 of III-NH)

E = 22,000 ksi at 1100F (Table TM-1 of II-D)

E = 22,400 ksi at 1050F (Table TM-1 of II-D)

Sm = 13,600 psi at 1050F (Table I-14.3A of III-NH)

Smt = 8500 psi at 1050F and at 130,000 hours (Table I-14.3A of III-NH)

Sm = 20,000 psi at 100F (Table 2A of II-D)

Sy = 30,000 psi at 100F (Table Y-1 of II-D)

Sy = 15,200 psi at 1050F (Table I-14.5 of III-NH)

Average Sy during the cycle = 22,600 psi

St = 12,200 psi at 1050F at 104 hours (Table I-14.3E of III-NH)

St = 8490 psi at 1050F at 130,000 hours (Table I-14.3E of III-NH)

= 10.4 10-6 in./in.-F at 1050F (Table TE-1 of II-D)

m = 0.3

Solution

Assumptions

Calculations in this example are based on thin shell equations to keep the calculations as

simple as possible. This was done to demonstrate and highlight the method of creep analysis.

For thick shells, Lames equations must be used.

The centerlines of the thin and thick shells are aligned.

The following calculations are based on the criteria of Fig. 4.8.

ASM_Jawad_Ch05.indd

132

MTC

02/24/2009 03:56AM

TABLE 5.2

PRIMARY MEMBRANE STRESS, PSI

Stress1

9608

4804

158

4725

2363

158

Pm

PmL

Pmr

Ri = 30.5 in.

3

Ri = 30.0 in.

1

2

The design stress at the beginning of the cycle is zero. Membrane stress due to design pressure of

315 psi is shown in Table 5.2.

The maximum membrane equivalent stress factors, Pm, are obtained from Eq. (4.14) as

For 1-in. thick shell = 8455 psi

For 2-in. thick shell = 4230 psi

From Fig. 4.8, the value of Pm for both 1-in. and 2-in. shells is less than the design allowable stress

of So (9800 psi). Hence, load-controlled design stress limits are adequate. The quantities PL and Pb

are not applicable in this case.

(b) Stress due to operating condition using elastic analysis

Both load-controlled and strain-controlled limits must be satised. The load-controlled limits

will be evaluated rst in accordance with Fig. 4.8.

(b.1) Load-controlled limits

The operating stress at the beginning of the cycle is zero. Stress due to operating pressure of

300 psi is shown in Table 5.3.

The maximum membrane equivalent stress factors, Pm, are obtained from Eq. (4.14) as

For 1-in. thick shell = 8050 psi

For 2-in. thick shell = 4025 psi

From Fig. 4.8, the values of Pm for the 1-in. and 2-in. thick shells are within the allowable stress

of Smt (8500 psi at 1050F). PL and Pb are not applicable in this case.

(b.2) Strain-controlled limits

The pressure as well as the temperature stresses must be included in the calculations. For thin

shells, the governing equations for thermal stress is obtained from Appendix B as

TABLE 5.3

PRIMARY MEMBRANE STRESS, PSI

Stress1

9150

4575

-150

4500

2250

-150

Pm

PmL

Pmr

Ri = 30.5 in.

3

Ri = 30.0 in.

1

2

ASM_Jawad_Ch05.indd

133

MTC

02/24/2009 03:56AM

134

Chapter 5

TABLE 5.4

STRESSES AT INSIDE SURFACE

Stress

Stress due to

pressure, psi

Stress due to

temperature, psi

One-inch shell

Membrane circumferential stress, Pm

Membrane axial stress, PmL

Membrane radial stress, Pmr

Circumferential bending stress, Qb

Axial bending stress, QbL

Radial bending stress, Qbr

9,150

4,575

-150

0

0

-150

0

0

0

-6,655

-6,655

0

Two-inch shell

Membrane circumferential stress, Pm

Membrane axial stress, PmL

Membrane radial stress, Pmr

Circumferential bending stress, Qb

Axial bending stress, QbL

Radial bending stress, Qbr

4,500

2,250

-150

0

0

-150

0

0

0

-14,975

-14,975

0

VT

V1

E D 'T i

2(1 P )

The stresses due to pressure and temperature of the 1.0-in. and 2.0-in. shells are given in Tables 5.4

and 5.5. Table 5.4 shows the inside surface stresses, and Table 5.5 shows the outside surface stresses.

Maximum equivalent stress values from Eq. (4.14)

Note that the calculated stresses shown below are based on the VIII-2 denition of equivalent stress

given by as given by the Von Mises expression. III-NH uses stress intensities based on Tresscas maximum shear theory, which are more conservative than those in the VIII-2 equivalent stress theory. In

TABLE 5.5

STRESSES AT OUTSIDE SURFACE

Stress

Stress due to

pressure, psi

Stress due to

temperature, psi

One-inch shell

Membrane circumferential stress, Pm

Membrane axial stress, PmL

Membrane radial stress, Pmr

Circumferential bending stress, Qb

Axial bending stress, QbL

Radial bending stress, Qbr

9,150

4,575

-150

0

0

-150

0

0

0

6,655

6,655

0

Two-inch shell

Membrane circumferential stress, Pm

Membrane axial stress, PmL

Membrane radial stress, Pmr

Circumferential bending stress, Qb

Axial bending stress, QbL

Radial bending stress, Qbr

4,500

2,250

-150

0

0

150

0

0

0

14,975

14,975

0

ASM_Jawad_Ch05.indd

134

MTC

02/24/2009 03:56AM

this section, to be consistent with VIII-2 policies, the III-NH procedures are used in conjunction with

the VIII-2 equivalent stress denition.

1.0-in. shell at inside surface

Membrane equivalent stress due to pressure, PLm = 8050 psi

Bending equivalent stress due to pressure, Q = 150 psi

Membrane equivalent stress due to temperature, PLm = 0 psi

Bending equivalent stress due to temperature, Q = 6655 psi

1.0-in. shell at outside surface

Membrane equivalent stress due to pressure, PLm = 8050 psi

Bending equivalent stress due to pressure, Q = 150 psi

Membrane equivalent stress due to temperature, PLm = 0 psi

Bending equivalent stress due to temperature, Q = 6655 psi

2.0-in. shell at inside surface

Membrane equivalent stress due to pressure, PLm = 4025 psi

Bending equivalent stress due to pressure, Q = 150 psi

Membrane equivalent stress due to temperature, PLm = 0 psi

Bending equivalent stress due to temperature, Q = 14,975 psi

2.0-in. shell at outside surface

Membrane equivalent stress due to pressure, PLm = 4025 psi

Bending equivalent stress due to pressure, Q = 150 psi

Membrane equivalent stress due to temperature, PLm = 0 psi

Bending equivalent stress due to temperature, Q = 14,975 psi

Stress in 1.0-in. thick shell

From Eqs. (5.1) and (5.2),

X

Y

PL Pb K Sy

QSy

0356

0301

All three tests (A-1, A-2, and A-3) will be checked in this example to demonstrate their applicability. In actual practice, the designer uses one of these tests to satisfy the limits. If the test limit is

exceeded, then the other tests are evaluated or a new design is tried.

Test A-1

Sa is the lesser of

1.25St = 1.25(12,200) =15,250 psi

or

05 SyL SyH

0356 0301 d 15250 22600

0657 0675

ASM_Jawad_Ch05.indd

135

MTC

02/24/2009 03:56AM

Chapter 5

136

Test A-2

This test is applicable because the temperature at the cold side of the cycle is below that given

in Table 5.1. From Eq. (5.7),

0356 0301

0657 100

Test A-3

The following two criteria will be checked rst for this test. Use the lesser of

PL Pb Q 3Sm

(1)

PL Pb Q 3Sm

(2)

or

PL + Pb + Q at inside surface = 6995 psi

PL + Pb + Q at outside surface = 14,085 psi

14085 313600

14085 40800

14085 d >1513600 058490@

14085 24600

_

Because 3S m < 3Sm, Eq. (2) controls.

Next, the thermal ratchet will be checked. Since X = 0.356, the required value of Y is given by Eq.

(5.7)

Y

0288 d 1X

1356

281

Thus, the stresses are satisfactory; however, for these criteria to be applicable in accordance with

III-NH, the negligible creep criteria must also be satised. This requires from Eqs. (5.9) and (5.10)

that

Ti Tid d 01

and

Hi

d 02%

where

Ti = total time that the metal is at temperature Ti

Tid = maximum allowable time from Table I-14.6A of III-NH at a temperature Ti and stress of 1.5Sy

at a temperature, Ti

i = creep strain at a stress level of 1.25Sy for the time T i

For 1.5Sy = 22,800 psi and Ti = 1050F, T id = 2000 hours and

ASM_Jawad_Ch05.indd

136

MTC

02/24/2009 03:56AM

(Tir ) = 130,000/2000 = 65 0.1

(Ti )

For 1.25Sy = 18,800 psi and T i = 1050F, from the isochronous stress-strain curve Fig. T-1800-A-6

Thus, although the creep modied shakedown and thermal ratchet limits are satised, the potential

for creep effects is signicant and test A-3 is not applicable. (Note that it would be more expeditious

in normal practice to check for negligible creep rst, before proceeding to stress evaluation.) The negligible creep criteria of test A-3 are conservative in that they assume a worst-case scenario wherein the

local stress is at the nominal ow stress.

Stress in 2.0-in. thick shell

From Eqs. (5.1) and (5.2),

X

PL Pb K Sy

QSy

1482522600

0178

0656

Test A-1

0178 0656

0835

0835 ! 0675

Test A-2

This test is applicable because the temperature at the cold side of the cycle is below that given

in Table 5.5. From Eq. (5.7),

0178 0656

0835 100

Test A-3

This test does not apply because the creep is not negligible.

5.4

The simplied inelastic analysis is based on the concept of requiring membrane stress in the core

of a cross-section to remain elastic, whereas bending stress is allowed to extend in the plastic region.

This concept is based on the Bree diagram (Bree, 1967, 1968). The Bree diagram assumes secondary

stress to be mainly generated by thermal gradients. The diagram (Fig. 5.3) is plotted with primary

stress as the abscissa and secondary stress as the ordinate. The diagram is divided into various zones

that dene specic stress behavior of the shell. It assumes an axisymmetric thin shell with an axisymmetric loading. It also assumes the thermal stress to be linear across the thickness. The actual derivation and assumptions made in constructing the Bree diagram are detailed in Appendix A. The various

limitations and zones of the diagram are as follows.

Limitations

It is assumed that the material has an elastic perfectly-plastic stress-strain diagram.

Because mechanical stress is considered primary stress, it cannot exceed the yield stress value

of the material. Thermal stress, on the other hand, is considered secondary stress and can thus

exceed the yield stress of the material.

ASM_Jawad_Ch05.indd

137

MTC

02/24/2009 03:56AM

138

Chapter 5

FIG. 5.3

BREE DIAGRAM (BREE, 1967)

Initial evaluation of the mechanical and thermal stresses in the elastic and plastic regions was

made without any consideration to relaxation or creep.

Final results were subsequently evaluated for relaxation and creep effect.

It is assumed that stress due to pressure is held constant, while the thermal stress is cycled.

Hence, pressure and temperature stress exist at the beginning of the rst half of the cycle and

only pressure exists at the end of the second half of the cycle.

Zone E

This zone is bounded by axis lines AB and AC as well as line BC, which is dened by the equation X + Y = 1.0.

Stress is elastic below the creep range.

ASM_Jawad_Ch05.indd

138

MTC

02/24/2009 03:56AM

Stress redistributes to elastic value above the creep range.

Zone S1

This zone is bounded by axis line CD as well as line BC (dened by the equation X + Y =

1.0), line DF (dened by the equation Y = 2), and line BF (dened by the equation X + Y/4 =

1.0).

Below the creep range, the stress is plastic on the outside surface of shell during the rst half

of rst cycle. The stress shakes down to elastic in all subsequent cycles.

Ratcheting does not occur below the creep range.

Ratcheting occurs in the creep range.

Shakedown is not possible at the creep range.

Zone S2

This zone is bounded by the axis line CD, line DF (expressed by the equation Y = 2), and line

FC (dened by equation Y(1 - X) = 1.0).

Zone S2 is a subset of zone S1.

Below the creep range, stress is plastic on both surfaces of shell during the rst half of rst

cycle. Stress shakes down to elastic in all subsequent cycles.

Ratcheting does not occur below the creep range.

Ratcheting occurs in the creep range.

Shakedown is not possible at the creep range.

Zone P

This zone is bounded by axis line DI, line DF (expressed by Y = 2), and line FG (expressed by

the equation XY = 1.0).

In this zone, alternating plasticity occurs in each cycle below the creep range.

Shakedown is not possible below as well as in the creep range.

Failure occurs due to low cycle fatigue below the creep range.

Shakedown is not possible at the creep range.

Zone R1

This zone is bounded by line BJ, which is also the Y axis, line BF (expressed as X + Y/4 = 1.0),

and line FH (expressed by the equation Y(1 - X) = 1.0).

Ratcheting occurs below as well as in creep range.

Shakedown is not possible below as well as in the creep range.

Zone R2

This zone is bounded by line FG (expressed as XY = 1.0), and line FH (expressed as Y

(1 - X ) = 1.0).

Ratcheting occurs below as well as in creep range.

Shakedown is not possible below as well as in the creep range.

The actual diagram used by ASME is shown in Fig. 5.4. The gure includes Z lines of constant

elastic core stress values (ODonnell and Porowski, 1974). The key feature of the ODonnell/Porowski

technique is identifying an elastic core in a component subjected to primary loads and cyclic secondary loads. Once the magnitude of this elastic core has been established, the deformation of the component can be bounded by noting that the elastic core stress governs the net deformation of the section.

Deformation in the ratcheting, R, regions of the Bree diagram can also be estimated by considering

individual cyclic deformation. The ASME simplied inelastic analysis procedure for strain limits consists of satisfying test B-1, test B-2, or test B-3. The following is a summary of these tests.

ASM_Jawad_Ch05.indd

139

MTC

02/24/2009 03:56AM

140

Chapter 5

FIG. 5.4

EFFECTIVE CREEP STRESS PARAMETER Z FOR SIMPLIFIED INELASTIC ANALYSIS

USING TEST NUMBERS B-1 AND B-3 (ASME, III-NH)

ASM_Jawad_Ch05.indd

140

MTC

02/24/2009 03:56AM

X includes all membrane, primary bending, and secondary bending stresses due to pressureinduced as well as thermal induced membrane stresses.

Y includes all thermal secondary stresses.

SyL is the yield stress at the cold end of the cycle.

5.4.1

1. The average wall temperature at one of the stress extremes dening each secondary equivalent

stress range, Q, is below the applicable temperature in Table 5.1.

2. The individual cycle cannot be split into sub-cycles.

3. Pressure-induced membrane and bending stresses and thermal induced membrane stresses are

classied as primary stresses for purposes of this evaluation.

4. Denitions of X and Y in Eqs. (5.4) and (5.5) apply for these two tests except that the value of

Sy as dened in these two equations is replaced with SyL, which is yield stress at the lower end of

the cycle.

5. These tests are applicable only in regimes E, S1, S2, and P in Fig. 5.4.

5.4.2

Test B-1

1. The peak stress is negligible.

2. c is less than the hot yield stress, SyH.

The procedure for applying test B-1 consists of the following steps:

1. Determine Z values from either Fig. 5.4 or the following values

Z

in region E

12

Y 1 2 >1 X Y @

in region S1

XY

in region S2 and P

(5.11)

Vc

ZSyL

(5.12)

4. Calculate the creep ratcheting strain from and isochronous curve using the quantity 1.25c and

the total hours for whole life.

5. The resulting strain should be less than 1% for the parent material and % for welded material.

5.4.3

Test B-2

This test is applicable to any structure and loading. However, Fig. 5.5, which takes into consideration peak stresses, is applicable in lieu of Fig. 5.4. The procedure for applying test B-2 consists of the

following steps:

1. Determine Z values from Fig. 5.5.

2. Calculate the effective creep stress c from

ASM_Jawad_Ch05.indd

141

MTC

02/24/2009 03:56AM

Chapter 5

142

FIG. 5.5

EFFECTIVE CREEP STRESS PARAMETER Z FOR SIMPLIFIED INELASTIC ANALYSIS

USING TEST NUMBERS B-2 (ASME, III-NH)

Vc

ZSyL

(5.13)

4. Calculate the creep ratcheting strain from and isochronous curve using the quantity 1.25c and

the total hours for whole life.

5. The resulting strain should be less than 1% for the parent material and % for weld material.

5.4.4

Test B-3

This test is used in regimes R1 and R2 in Fig. 5.4. It can also be used in regimes S1, S2, and P to

minimize the conservatism of tests B-1 and B-2. The following conditions are applicable to this test:

1. Applies to axisymmetric structures subjected to axisymmetric loading and away from local structural discontinuities where peak stress is negligible.

2. When peak stress is negligible, the constraint on local structural discontinuities may be considered satised.

3. Wall membrane forces from overall bending of a vessel can be conservatively included as axisymmetrical forces.

ASM_Jawad_Ch05.indd

142

MTC

02/24/2009 03:56AM

4. The denitions of X and Y in Eqs. (5.4) and (5.5) apply, but XL, YL, XH, and YH are calculated for

the cold and hot ends using SyL and SyH, respectively.

The procedure for this test consists of calculating the inelastic strains, , in accordance with the

equation

Q + K + G

(5.14)

where

S = total inelastic strain accumulated in the lifetime of the component

S = enhanced creep strain increments due to relaxation of c stresses

S = plastic ratchet strain increments for cycles in regimes S1, S2, P, R1, and R

S = inelastic strains obtained from the isochronous curves as in test B-1, ignoring the increase in

c stress for cycles evaluated in this section and detailed inelastic analyses.

Values of and are obtained as follows.

Values of

Plastic ratcheting occurs in cycles when cL > SyH. The increment of plastic ratchet within

a given cycle is

K

for Z L 10

(5.15)

and

K

for Z L ! 10

(5.16)

Equation (5.15) must be used only at the hot extreme in regimes S1, S2, and P. Equation

(5.16) must be used at both extremes in regimes R1 and R2.

Values of

For cycles where cL > SyH, the enhanced creep strain increment due to stress relaxation is

G

2

2

1 S yH V c

Vc

EH

(5.17)

For cycles where cL < SyH, the enhanced creep strain increment due to stress relaxation is

G

2

Vc

1 V cL

EH

(5.18)

Vc

Example 5.2

Use the data from Example 5.1 to evaluate stress away from any discontinuities due to operating

condition using simplied inelastic analysis. Use the isochronous curves shown in Fig. 5.6.

Solution

Normally, there is no need to use simplied inelastic analysis if the elastic analysis is satised.

However, for this example, the simplied inelastic analysis is performed to compare the results of

the two methods.

Tests B-1 and B2

Calculations for X and Y are based on using the yield stress at the ambient temperature. From

Eqs. (5.1) and (5.2),

ASM_Jawad_Ch05.indd

143

MTC

02/24/2009 03:56AM

144

Chapter 5

FIG. 5.6

ISOCHRONOUS STRESS-STRAIN CURVES FOR 304 STAINLESS STEEL (ASME, III-NH)

One-inch shell

Test B-1

Pm Sy

[ membrane stress due to pressure (bending stress due to pressure

membrane stress due to temperature] SyL

ASM_Jawad_Ch05.indd

144

MTC

02/24/2009 03:56AM

circumferential Pm = 9150 + (0)/1.25 + 0 = 9510 psi

axial Pm = 4175 + (0)/1.25 + 0 = 4175 psi

radial Pm = -150 + (-150)/1.25 + 0 = -270 psi

X = 8160/30,000 = 0.272

Y

QSy

(bending stress due to temperature)/S yL

6655 30000

0272

Z

0272

SyL

Vc

30000 psi

From the isochronous curves (Fig. 5.6), with stress = 1.25c (10,200 psi) and expected life of

130,000 hours, we obtain a strain of 0.42%. This value is acceptable because it is less than the

permissible value of 1%.

Tests B-2 and B-3

These tests need not be performed because test B-1 is satised.

Two-inch shell

Test B-1

X Pm Sy

[membrane stress due to pressure (bending stress due to pressure

membrane stress due to temperature] SyL

circumferential Pm = 4500 + (0)/1.25 + 0 = 4500 psi

axial Pm = 2250 + (0)/1.25 + 0 = 2250 psi

radial Pm = -150 + (-150)/1.25 + 0 = -270 psi

X = 4130/30,000 = 0.138

Y QSy

(bending stress due to temperature)/S yL 1497530000 0500

ASM_Jawad_Ch05.indd

145

MTC

02/24/2009 03:56AM

146

Chapter 5

Z

0138

SyL

Vc

0272 30000

30000 psi

8160 psi

From the isochronous curves (Fig. 5.5), with stress = 1.25c (5175 psi) and expected life of

130,000 hours, we obtain a strain of 0.026%. This value is acceptable because it is less than

the permissible value of 1%.

Tests B-2 and B-3

These tests need not be performed because test B-1 is satised.

It should be pointed out at this time that the expected design life of 130,000 hours for this

shell is longer than the life of 100,000 hours set for the allowable stress criterion for I and VIII.

Also, the allowable stress of 8500 psi in III-NH is conservatively based on a 67% criterion

rather than the 80% set for I and VIII. Additionally, the rules of III-NH for operating condition have to be complied with once the temperature is in the creep range. Such criterion is not

mandatory in I or VIII below the creep range.

Example 5.3

Check the shell stress due to operating condition in accordance with III-NH strain-controlled limits

in Example 4.5 of Chapter 4. The following values are obtained from III-NH:

Sy = 30 ksi at room temperature

Sy = 26.52 ksi at 810F

St = 17,200 psi for 100,000 hours

Smt = 17,120 psi

Solution

The equivalent stresses calculated in Example 4.5 are

Stress, psi

Shell

Head

PL

PL + Pb

PL + Pb/Kt

Q

7,315

7,315

7,315

27,540

425

910

810

225

Shell

Elastic analysis

X = 7315/28,260 = 0.259, Y = 27,540/28,260 = 0.975

This test, which is more conservative than Test A-2, does not apply because one temperature

extreme of the stress cycle is below the creep threshold temperature for applicability of Test A-2.

(b) Test A-2

X + Y = 1.233

ASM_Jawad_Ch05.indd

146

MTC

02/24/2009 03:56AM

(c) Test A-3

This test does not apply because the negligible creep criteria for Test A-3 are not satised.

Simplied inelastic analysis

Test B-1

X = Pm/Sy = [PL + (Pb + Q)/Kt]/SyL

circumferential Pm = 0 + (0 + 9240)/1.25 = 7390 psi

axial Pm = 7200 + (0 + 30800)/1.25 = 31840 psi

radial Pm = -225 + (0 - 225)/1.25 + 0 = -405 psi

X = 29,135/30,000 = 0.138

Y = Q/Sy

= (bending stress due to temperature)/SyL = 0/30,000 = 0

Z = 0.971

SyL= at the cold end of the cycle is at room temperature = 30,000 psi

c = 0.971(30,000) = 29,135 psi

However, Test B-1 also requires that c SyH. In this case, SyH = 26,520 psi which is less than c

and Test B-1 is, thus, not applicable. Test B-2 is more conservative than Test B-1 and the applicability of Test B-3 to cases where cyclic stresses are solely due to pressure is not clearly dened. Thus

the remaining alternatives are to consider a thicker section or resort to inelastic analysis.

Problems

5.1 A hydrotreater has an inside diameter of 12 ft and a length of 50 ft. The design pressure is

2400 psi and the design temperature is 975F. The operating pressure is 2300 psi and the operating temperature is 950F. The material of construction is 2.25Cr-1Mo steel. Expected life

is 200,000 hours.

(a) Determine the required thickness in accordance with VIII-1.

(b) Evaluate the load-controlled limits.

(c) Evaluate the strain- and deformation-controlled limits using

i. Elastic tests A

ii. Simplied inelastic tests B

Data:

So = 9400 psi at 975F (Table I-14.2 of III-NH)

Eo = 1.0

E = 24,930 ksi at 975F (Table TM-1 of II-D)

E = 25,150 ksi at 950F (Table TM-1 of II-D)

ASM_Jawad_Ch05.indd

147

MTC

02/24/2009 03:56AM

148

Chapter 5

Smt = 7850 psi at 950F and at 200,000 hours (Table I-14.3D of III-NH)

Sm = 20,000 psi at 100F (Table 2A of II-D)

Sy = 30,000 psi at 100F (Table Y-1 of II-D)

Sy = 24,800 psi at 950F (Table Y-1 of II-D)

Average Sy during the cycle = 27,400 psi

St = 11,300 psi at 950F at 104 hours (Table I-14.4D of III-NH)

St = 7850 psi at 950F at 200,000 hours (Table I-14.4D of III-NH)

5.2

A boiler header has an outside diameter of 20 in. and a length of 20 ft. The design pressure

is 2900 psi and the design temperature is 1000F. The operating pressure is 2700 psi and the

operating temperature is 975F. The material of construction is 2.25Cr-1Mo steel. Expected

life is 300,000 hours.

(a) Determine the required thickness in accordance with ASME-I.

(b) Evaluate the load-controlled limits.

(c) Evaluate the strain- and deformation-controlled limits using

i. Elastic tests A

ii. Simplied inelastic tests B

Data:

So = 8000 psi at 1000F (Table I-14.2 of III-NH)

Eo = 0.65 (ligament efciency) for circumferential stress calculations

Eo = 0.95 (ligament efciency) for longitudinal stress calculations

E = 24,700 ksi at 1000F (Table TM-1 of II-D)

E = 24,930 ksi at 975F (Table TM-1 of II-D)

Sm = 16,000 psi at 975F (Table I-14.3D of III-NH)

Smt = 6250 psi at 975F and at 300,000 hours (Table I-14.3D of III-NH)

Sm = 20,000 psi at 100F (Table 2A of II-D)

Sy = 30,000 psi at 100F (Table Y-1 of II-D)

Sy = 24,250 psi at 975F (Table Y-1 of II-D)

Average Sy during the cycle = 27,400 psi

St = 10,000 psi at 975F at 104 hours (Table I-14.4D of III-NH)

St = 6250 psi at 975F at 300,000 hours (Table I-14.4D of III-NH)

ASM_Jawad_Ch05.indd

148

MTC

02/24/2009 03:56AM

ST. LOUIS, MISSOURI)

ASM_Jawad_Ch06.indd

150

MTC

02/24/2009 04:22AM

CHAPTER

6

CREEP-FATIGUE ANALYSIS

6.1

INTRODUCTION

In this chapter, the American Society of Mechanical Engineers (ASME) criterion for designing

pressure vessels under repetitive cyclic loading conditions in the creep range is discussed. Such cycles

are generally encountered in power plants as well as petrochemical plants under normal operating

conditions. The assumption of repetitive cycles enables us to focus rst on explaining the ASME procedure for designing components under cyclic loading in the creep range without having to deal with

the complex problem of variable cyclic conditions. Variable cyclic loading due to upset, regeneration,

or other emergency conditions are discussed later in this chapter.

Cyclic analysis is straightforward at temperatures below the creep range. The maximum calculated

stresses are compared to fatigue curves obtained from experimental data with an appropriate factor

of safety. The fatigue curves take into consideration such factors as average versus minimum stress

values, effect of mean stress on fatigue life, and size effects.

In the creep range, cyclic life becomes more difcult to evaluate (Jetter, 2002). Stress relaxation at a

given point affects the cyclic life of a component. The level of triaxiality and stress concentration factors play a signicant role on creep-fatigue life at elevated temperatures and Poissons ratio needs to

be adjusted to account for inelastic stress levels. In addition, fatigue strength tends to decrease (Frost

et. al., 1974) with an increase in temperature due to surface oxidation or chemical attack. These and

other factors contribute to the tediousness and complexity of creep-fatigue evaluation.

The data required to evaluate the cyclic stress in a given material is extensive. Data needed for

creep-rupture analysis at various temperatures include stress-strain diagrams, yield stress and tensile

strength, creep and rupture data, modulus of elasticity, and isochronous curves. Large amounts of

time and cost are involved in obtaining such data. As a consequence, data for creep analysis has been

developed for only ve materials (III-NH): 2.25Cr-1Mo and 9Cr steels, 304 and 316 stainless steels,

and 800H nickel alloy.

The ASME rules for cyclic loads in the creep range consist of determining points in a cycle time

where the stresses at a given location in a vessel are at a maximum level. The stresses are then checked

against limiting values for fatigue and creep. The analysis for cyclic loading consists of evaluating

stresses based on load-controlled as well as strain-controlled limits similar to the procedure discussed

in Chapters 4 and 5.

6.2

The rules for creep-fatigue evaluation discussed in this section are applicable when

1. The rules of Section 5.4 for tests A-1 through A-3 are met and/or the rules of Section 5.5 for

tests B-1 and B-2 with Z < 1.0 are met. However, the contribution of stress due to radial thermal

151

ASM_Jawad_Ch06.indd

151

MTC

02/24/2009 04:22AM

152

Chapter 6

gradients to the secondary stress range may be excluded for this assessment of the applicability

of elastic creep-fatigue rules, A-1 and A-2.

2. The (PL + Pb + Q) 3Sm rule is met using for 3Sm the lesser of (3Sm) and (3S m) as dened in

test A-3.

3. Pressure-induced membrane and bending stresses and thermal induced membrane stresses are

classied as primary (load-controlled) stresses.

The analysis procedure is performed by using the following six steps.

Step 1

Determine the total amount of hours, TH, expended at temperatures in the creep range.

Step 2

Dene the hold temperature, THT, to be equal to the local metal temperature that occurs during

sustained normal operation.

Step 3

Unless otherwise specied, for each cycle type j, dene the average cycle time as

Tj = TH/(nc)j

(6.1)

where

(nc)j = specied number of applied repetitions of cycle type j

TH = total number of hours at elevated temperatures for the entire service life as dened in step 1

Step 4

A modied strain, Dmod, is calculated in this step. The procedure consists of calculating rst a

maximum strain, Dmax, and then modifying it to include the effect of stress concentration factors. The

maximum elastic strain range during the cycle is calculated as

max = 2Salt/E

(6.2)

where

E

= modulus of elasticity at the maximum metal temperature experienced during the cycle

2Salt = maximum stress range during the cycle excluding geometric stress concentrations = PL +

Pb + Q

Dmax = maximum equivalent strain range

The maximum elastic strain range is then used to calculate a modied strain (Dmod) that includes

the effect of local plasticity and creep, which can signicantly increase the strain range at stress concentrations. Subsection NH gives the designer the option of calculating this quantity by using any

one of three different methods. These methods, of varying complexity and conservatism, are based

on modications of the Neuber equation (Neuber, 1961). Neubers basic equation (Bannantine et al.,

1990) is of the form

K = (KK)1/2

(6.3)

K 2Se =

(6.4)

ASM_Jawad_Ch06.indd

152

MTC

02/24/2009 04:22AM

where

e = strain away from concentration

K = theoretical stress concentration

K = stress concentration due to strain

K = stress concentration due to stress

S = stress away from concentration

= strain at concentration

= stress at concentration

Equation (6.3) indicates that the total stress concentration at a point in the plastic or creep region

is equal to the square root of the products K and K. The value of K decreases, whereas that of K

increases with an increase in yield and creep levels. Equation (6.4), which is an alternate form of Eq.

(6.3), shows that the total stress concentration is a function of the product of the actual strain and

actual stress at a given point.

All three methods (Severud, 1987 and 1991) of calculating modied strain, Dmod, use a composite

stress-strain curve as shown by Fig. 6.1. The composite stress-strain curve is constructed by adding

the elastic stress-strain curve for the stress range SrH to the appropriate time-independent (hot tensile)

isochronous stress-strain curve for the material at a given temperature. The value of SrH can be conservatively assumed as equal to 0.5St. The conceptual basis for extending the elastic range is shown in

Fig. 1.21, illustrating how the elastic stress range is extended by an amount given by the hot relaxation

FIG. 6.1

STRESS-STRAIN RELATIONSHIP (ASME, III-NH)

ASM_Jawad_Ch06.indd

153

MTC

02/24/2009 04:22AM

154

Chapter 6

First method

The governing equation is

(6.5)

where

= either the equivalent stress concentration factor, as determined by test or analysis, or the

maximum value of the theoretical elastic stress concentration factor in any direction for

the local area under consideration1

*

S

= stress indicator determined by entering the composite stress-stain curve of Fig. 6.1 at a

strain range2 of (Dmax)

S

= stress indicator determined by entering the composite stress-stain curve of Fig. 6.1 at a

strain range2 of (KscDmax)

Dmod = modied maximum equivalent strain range that accounts for the effects of local plasticity

and creep

Ksc

Second method

Dmod is calculated from the equation

Dmod = KeKscDmax

(6.6)

where

Ke = 1.0

if KscDmax < (3S m)/E

_

Third method

Dmod is given by

Dmod = S *Ksc2 Dmax/Dmod

(6.7)

where

S*

= stress obtained from an isochronous curve at a given value of Dmax

Dmod = range of effective stress that corresponds to the strain range Dmod2

Step 5

In this step, a stress, Sj, is obtained. It corresponds to a strain value that includes elastic, plastic,

and creep considerations. Once the quantity Dmod is known, then the total strain range, t, is obtained

from the following equation

t = KvDmod + KscDc

(6.8)

where

t = total strain range

1

The equivalent stress concentration factor is dened as the effective (von Mises) primary plus secondary plus peak stress

divided by the effective primary plus secondary stress. Note that fatigue strength reduction factors developed from lowtemperature continuous cycling fatigue tests may not be acceptable for dening Ksc when creep effects are not negligible.

2

Both Dmod and Dmod in Eq. (6.7) are unknown, and they must be solved graphically by curve tting the appropriate composite stress-strain curve. For this reason, most designers opt to use either Eq. (6.5) or Eq. (6.6) because they are easier to solve.

ASM_Jawad_Ch06.indd

154

MTC

02/24/2009 04:22AM

and

Kv = 1.0 + f (Kv - 1.0), but not less than 1.0

(6.9)

where

f = triaxiality factor obtained from Fig. 6.23

Kv = plastic Poisson ratio adjustment factor obtained from Fig. 6.34

The creep strain increment, c, is obtained from an isochronous stress-strain curve similar to the

one shown in Fig. 5.6. The stress value for entering the gure is obtained from the quantity 1.25c,

where c is obtained from Section 5.5. The time used in the gure for determining c is obtained by

one of two methods as follows:

Method (1): the time based on one cycle

Method (2): the time based on the total number of hours during the life of the component and the

resultant strain is then divided by the number of cycles

Method (1) is generally applicable to components with a small number of cycles and high membrane stress such as hydrotreaters.

Method (2) is generally applicable to components with repetitive cycles and small membrane stress

such as headers in heat recovery steam generators.

Finally, a value of Sj is obtained from an appropriate isochronous chart using Eq. (6.8) for strain and

the time-independent curve in the chart. Sj is dened as the initial stress level for a given cycle.

Step 6

The relaxation stress, S r , during a given cycle is evaluated in this step. Two methods are provided

for this evaluation. The rst method requires an analytical estimate of the uniaxial stress relaxation

adjusted with correction factors to account for the retarding effects of multiaxiality and elastic followup. The adjusted relaxed stress level Sr is thus a function of the initial stress determined from t, the

analytically determined uniaxial relaxed stress, and a factor 0.8G accounting for elastic follow-up and

multiaxiality. The equation is expressed as

Sr = Sj - 0.8G(Sj -S r)

(6.10)

where

G = the smallest value of the multiaxiality factor as determined for the stress state at each of the two

extremes of the stress cycle. The multiaxiality factor is dened as:

>V 1 05 V 2 V 3@

>V1 03 V 2 V 3 @

1, 2, and 3 are principal stresses, exclusive of local geometric stress concentration factors, at

the extremes of the stress cycle, and are dened by

|1| |2| |3|

3

4

This gure is based on the relationship between strain range and shakedown criteria.

ASM_Jawad_Ch06.indd

155

MTC

02/24/2009 04:22AM

156

Chapter 6

FIG. 6.2

INELASTIC MULTIAXIAL ADJUSTMENTS (ASME, III-NH)

FIG. 6.3

ADJUSTMENT FOR INELASTIC BIAXIAL POISSONS RATIO (ASME, III-NH)

ASM_Jawad_Ch06.indd

156

MTC

02/24/2009 04:22AM

Sr = relaxed stress level at time T adjusted for the multiaxial stress state

The second method for determining stress relaxation is based on the isochronous stress strain

curves. Starting at the stress level determined from t, the time to relax to lower stress levels is determined by moving vertically down at a constant strain until intercepting the curve for the time of

interest as shown in Fig. 6.4. Because of the conservatism inherent in this approach (Severud, 1991),

multiaxial and elastic follow-up corrections are not required.

The stress in either the adjusted analytical relaxation or that obtained from the isochronous

curves is not allowed to relax below a factor of 1.25 times the elastic core stress, c, as determined by

the procedures for evaluation of the strain limits using simplied inelastic analysis. This lower stress

value, SLB, is illustrated in Fig. 6.5.

Step 7

The governing equation for creep fatigue (Curran, 1976) is given by

[(nc/Nd)j + (D/Td)k] < Dcf

(6.11)

where

Dcf = total creep-fatigue damage factor obtained from Fig. 6.6

K

= 0.90

(Nd)j = number of design allowable cycles for cycle type, j, obtained from a design fatigue data using

the maximum strain value during the cycle

(nc)j = number of applied repetitions of cycle type, j

FIG. 6.4

A METHOD OF DETERMINING RELAXATION (ASME, III-NH)

ASM_Jawad_Ch06.indd

157

MTC

02/24/2009 04:22AM

158

Chapter 6

FIG. 6.5

STRESS-RELAXATION LIMITS FOR CREEP DAMAGE (ASME, III-NH)

FIG. 6.6

CREEP-FATIGUE DAMAGE ENVELOPE (ASME, III-NH)

ASM_Jawad_Ch06.indd

158

MTC

02/24/2009 04:22AM

(Td)k = allowable time duration determined from stress-to-rupture data for a given stress, (Sr/K)

for base material and [Sr/(KR)] for weldments, and the maximum temperature at the point

of interest and occurring during the time interval, k

(D)k = duration of the time interval, k

R

= weld strength reduction factor from Table I-14.10 of III-NH.

The rst term of Eq. (6.11) pertains to cyclic loading, whereas the second term pertains to creep

duration. A substantial amount of calculations is required before solving Eq. (6.11). The solution can

be made using either elastic analysis or inelastic analysis. In this section, elastic analysis, which is the

easiest to perform but the most conservative, is presented.

It should be noted that the quantity (D/Td)k in Eq. (6.11) is determined by one of two methods

as follows:

Method (1): When the total strain range is t > (3S m )/E, then the time interval k for stress relaxation is based on one cycle and the result is multiplied by the total number of cycles during

the life of the component.

Method (2): When the total strain range is t (3S m )/E, then shakedown occurs and the time

interval k for stress relaxation can be dened as a single cycle for the entire design life. When

applicable, method (2) will yield a lower calculated creep damage than method (1).

Example 6.1

Use the data from Examples 5.1 and 5.2 to evaluate the life cycle of the 2.0-in. thick shell for 130,000

hours (15 years) due to creep and fatigue conditions. Assume an arbitrary stress concentration factor,

Ksc, of 1.10 for the longitudinal weld conguration. The design fatigue strain range values are shown

in Table 6.1 and the stress to rupture values are given in Table 6.2.

TABLE 6.1

DESIGN FATIGUE STRAIN RANGE FOR 304 STAINLESS STEEL (ASME, III-NH)

t, strain range (in./in.) at

Number of

cycles, Nd

10

20

40

102

2 102

4 102

103

2 103

4 103

104

2 104

4 104

105

2 105

4 105

106

100F

800F

900F

0.051

0.036

0.0263

0.018

0.0142

0.0113

0.00845

0.0067

0.00545

0.0043

0.0037

0.0032

0.00272

0.0024

0.00215

0.0019

0.050

0.0345

0.0246

0.0164

0.0125

0.00965

0.00725

0.0059

0.00485

0.00385

0.0033

0.00287

0.00242

0.00215

0.00192

0.00169

0.0465

0.0315

0.0222

0.0146

0.011

0.00845

0.0063

0.0051

0.0042

0.00335

0.0029

0.00254

0.00213

0.0019

0.0017

0.00149

1000F

1100F

1200F

1300F

0.0335

0.0217

0.0146

0.0093

0.0069

0.00525

0.00385

0.00315

0.00263

0.00215

0.00187

0.00162

0.00140

0.00123

0.0011

0.00098

0.0297

0.0186

0.0123

0.0077

0.0057

0.00443

0.00333

0.00276

0.0023

0.00185

0.00158

0.00138

0.00117

0.00105

0.00094

0.00084

ASM_Jawad_Ch06.indd

159

0.0425

0.0284

0.0197

0.0128

0.0096

0.00735

0.0055

0.0045

0.00373

0.00298

0.00256

0.00224

0.00188

0.00167

0.0015

0.0013

0.0382

0.025

0.017

0.011

0.0082

0.0063

0.0047

0.0038

0.0032

0.0026

0.00226

0.00197

0.00164

0.00145

0.0013

0.00112

MTC

02/24/2009 04:22AM

160

Chapter 6

TABLE 6.2

MINIMUM STRESS-TO-RUPTURE VALUES FOR 304 STAINLESS STEEL (ASME, III-NH)

U.S. customary units

Temp.

(F)

1 hr

10

hr

30

hr

102

hr

3 102

hr

103

hr

3 103

hr

104

hr

3 104

hr

105

hr

3 105

hr

800

850

900

950

1000

1050

1100

1150

1200

1250

1300

1350

1400

1450

1500

57

56.5

55.5

54.2

52.5

50

45

38

32

27

23

19.5

16.5

14.0

12.0

57

56.5

55.5

54.2

50

41.9

35.2

29.5

24.7

20.7

17.4

14.6

12.1

10.2

8.6

57

56.5

55.5

51

44.5

37

31

26

21.5

17.9

15

12.6

10.3

8.8

7.2

57

56.5

55.5

48.1

39.8

32.9

27.2

22.5

18.6

15.4

12.7

10.6

8.8

7.3

6.0

57

56.5

51.5

43

35

28.9

23.9

19.3

15.9

13

10.5

8.8

7.2

5.8

4.9

57

56.5

46.9

38.0

30.9

25.0

20.3

16.5

13.4

10.8

8.8

7.2

5.8

4.6

3.8

57

50.2

41.2

33.5

26.5

21.6

17.3

13.9

11.1

8.9

7.2

5.8

4.7

3.8

3.0

57

45.4

36.1

28.8

22.9

18.2

14.5

11.6

9.2

7.3

5.8

4.6

3.7

2.9

2.4

51

40

31.5

24.9

19.7

15.5

12.3

9.6

7.6

6.0

4.8

3.8

3.0

2.3

1.8

44.3

34.7

27.2

21.2

16.6

13.0

10.2

8.0

6.2

4.9

3.8

3.0

2.3

1.8

1.4

39

30.5

24

18.3

14.9

11.0

8.6

6.6

5.0

4.0

3.1

2.4

1.9

1.4

1.1

Solution

Assumptions

The calculations in this example are based on thin shell equations to keep the calculations as

simple as possible. This is done to demonstrate and highlight the method of creep analysis. For

thick shells, Lames equations must be used.

The following calculations are based on the criteria in Fig. 4.8.

Ksc = 1.1.

To apply this methodology, it is necessary to rst verify that the prerequisites identied in Section

6.2 and repeated here have been satised:

1. The rules of Section 5.4 for Tests A-1 through A-3 are met and/or the rules of Section 5.5 for

Tests B-1 and B-2 with Z < 1.0 are met. However, the contribution of stress due to radial thermal

gradients to the secondary stress range may be excluded for this assessment in Tests A-1 and A-2

for the applicability of elastic creep-fatigue rules.

2. The (PL + Pb + Q) 3Sm rule is met using for 3Sm the lesser of (3Sm) and (S m) as dened in Test

A-3.

3. Pressure-induced membrane and bending stresses and thermal induced membrane stresses are

classied as primary (load-controlled) stresses.

Requirement # 1 is satised: Both A-2 and B-1 are satised.

Requirement # 2 is satised.

Requirement # 3 is satised.

Step 1

Total hours, TH = 130,000 hours (15 years)

ASM_Jawad_Ch06.indd

160

MTC

02/24/2009 04:22AM

Step 2

Hold temperature, THT = 1050F.

Step 3

n c j

j

cycle time T

number of cycles j

10

TH n c j

130000 10

A review of Tables 5.4 and 5.5 shows the combination of pressure and thermal stresses in the 2.0-in.

shell to be more severe than in the 1.0-in. shell. The tables also indicate that the membrane stress in

the 1.0-in. shell is more severe. The following creep-fatigue calculations are performed on the 2.0-in.

shell to illustrate the procedure. However, the designer must recognize that an analysis of the 1.0-in.

shell may result in a more severe condition due to the higher core stress.

The values obtained for the outside stress of the 2.0-in. shell given in Table 5.5 are summarized in

Table 6.3.

The following values are obtained from Table 6.3:

Total circumferential stress = 4500 +14,975 = 19,475 psi

Total longitudinal stress = 2250 + 14,975 = 17,225 psi

Total radial stress = -150 + 150 = 0

The equivalent Von Mises stress is obtained from Eq. (4.14) as

Se = 18,450 psi

It must be noted that the rules of III-NH use the Tresca criterion for the elastic analysis in steps 4

through 7 below. However, VIII-2 uses the Von Mises criterion that has less conservatism in it. The

following calculations are based on the Von Mises criterion that is deemed adequate for ASME I and

VIII applications.

Step 4

A modied strain, Dmod, is calculated in this step. The procedure consists of calculating a maximum

strain and then modifying it to include the effect of stress concentration factors. We start by calculating the maximum elastic strain range Dmax in accordance with Eq. (6.5).

TABLE 6.3

OUTSIDE STRESS FOR 2.0-in. SHELL

Two-inch shell

Stress due to

pressure, psi

Stress due to

temperature, psi

Membrane radial stress, Pmr

4,500

2,250

-150

0

0

150

0

0

0

14,975

14,975

0

ASM_Jawad_Ch06.indd

161

MTC

02/24/2009 04:22AM

162

Chapter 6

' H max

2Salt E

Se E

1845022400000

000082

Next, the value of modied strain Dmod is calculated. Step 4 gives the designer the choice of using

three different methods for calculating this quantity. The rst two methods, given by Eqs. (6.4) and

(6.5), will be used in order to compare the results.

First method for calculating modied strain

The value of S * is obtained from composite stress-strain curve (Fig. 6.1) with a strain value of

Dmax and the value of S is obtained from the same curve with a strain value of (Ksc)Dmax. However,

for this problem it is evident that the values of Dmax and (Ksc)Dmax are within the proportional limit

of Fig. 6.1. Under these conditions, S = (Ksc) S* and the equation for Dmod may be rewritten as

2

D e mod = S */S Ksc

D e max

2

= [S */(Ksc S *)] Ksc

D e max

= Ksc D e max

= 0.00090

The value of Ke in Eq. (6.6) must rst be calculated.

3Sm = 1.5Sm + St /2 = 1.5(13,600) + (8490)/2 = 24,645

3S m

E = 24,645/22,400,000 = 0.0011

Because (KscDmax) is less than (3S m )/E, the value of Ke is obtained from Eq. (6.6) as

Ke = 1.0

D e mod = Ke Ksc D e max

= (1.0)(1.1)(0.00082)

= 0.00090

This is identical to the result obtained with Method 1, which is the expected result within the

proportional limit.

Step 5

The total strain is obtained from Eq. (6.8)

t = KvDmod + KscDc

where

Kv = 1.0 + f (Kv - 1.0), but not less than 1.0

ASM_Jawad_Ch06.indd

162

MTC

02/24/2009 04:22AM

Kv = 1.0

The next step is to obtain the value of f from Fig. 6.2 by using the triaxiality factor (TF). From Table

6.3,

V1

4500 14976

V2

150 150

V3

2250 14976

19476 psi

0 psi

17226 psi

19476 0 17226

TF

0707 >19476

0 2

f

1 2

097

Kv = 1.0 + (0.97)(1.0 - 1.0) = 1.0

The elastic core stress, c, is obtained from Example 5.2 and is equal to 4140 psi. From the isochronous curves (Fig. 5.6), with stress of 1.25c = 1.25(4140) =5175 psi, a strain value of 0.00026 is

obtained. This strain value is on the elastic portion of the curve and is essentially independent of time

duration, at least up to the total design life of 130,000 hours. The ASME rules state that the creep strain

increment may be determined for each cycle, 13,000 hours in this example, or from the strain over the

total life divided by the number of cycles. In the latter case, the applicable strain, Dc, is equal to

Dc = 0.00026/10 = 0.000026

t(1.0)(0.00090) + (1.1)(0.000026) = 0.00093

Sj = 15,000 psi

Step 6

From Fig. 5.6, stress-versus-time values are obtained as shown in Table 6.4.

TABLE 6.4

STRESS-VERSUS-TIME VALUES

Time (hr)

Stress (ksi)

Sj

30

100

300

1,000

3,000

10,000

30,000

100,000

300,000

ASM_Jawad_Ch06.indd

163

15.0

13.7

12.0

10.5

9.0

8.0

7.7

7.2

6.8

6.6

MTC

02/24/2009 04:22AM

164

Chapter 6

A plot of these values, with a limit of 1.25c = (1.25)(4140) = 5175 psi, would look similar to Fig.

6.5.

Step 7

Equation (6.11) is solved in this step. The rst part of the equation, (nc/Nd)j, is obtained as

follows.

First part of Eq. (6.11)

nc = 10

The value of Nd is obtained from Table 6.1. A cycle life of >1,000,000 is obtained at a temperature

of 1050F and strain, t = 0.001.

(nc/Nd)j = 10/1,000,000 0.0

The second part of Eq. (6.11), (D/Td)k, is determined numerically from Table 6.4 and from the

differential equation dT/Td. The quantity (3S m )/E, calculated above, is equal to 0.0011, whereas the

value of t is equal to 0.00093. Because t (3S m )/E, shakedown occurs and the peak stress does not

reset on each cycle. Under this scenario, peak stress can be assumed to relax throughout the design life

and the cumulated creep damage is given by Table 6.4 at 130,000 hours. Calculations are summarized

in Table 6.5.

Column A in Table 6.5 gives the incremental locations for calculating stresses from Table 6.4.

Column B shows the stresses obtained from Table 6.4 and column H lists the incremental duration

assumed for these stresses. Column C adjusts the stress values by the K factor and column G lists the

number of hours obtained from Table 6.2 using the stress values in column C. Column I shows the

ratio of column H over column G, which is the numerical value of the quantity (dT/Td).

The 1050F line in Table 6.2 terminates at the 11,000 psi stress level. Stress levels in column C drop

below the 11,000 psi level in rows 6 through 10. The Larson-Miller parameter, PLM, is used to obtain

approximate rupture life by using the higher temperature levels given in Table 6.2. These calculations

are shown in columns D, E, and F. Column D lists the temperatures from Table 6.2 corresponding

to the stress levels in column C. Column E calculates the corresponding PLM from Eq. (1.3) for the

temperature shown in column D at 300,000 hours. Column F recalculates the hours corresponding to

a temperature of 1050F from Eq. (1.3) with the corresponding PLM factor.

TABLE 6.5

NUMERICAL CALCULATIONS OF (T/Td)k

A

Location

(hr)

Sj

30

100

300

1,000

3,000

10,000

30,000

100,000

Total

Sr(ksi)

Sr/0.9

(ksi)

Equivalent

temp. (F)

PLM

Time (hr)

Table 6.2

Table 6.4

H/G

15.0

13.7

12.0

10.5

9.0

8.0

7.7

7.2

6.8

16.67

15.22

13.33

11.67

10.00

8.89

8.56

8.00

7.56

673,700

1,657,600

2,176,400

3,745,500

5,749,100

21,300

37,800

90,800

263,000

673,700

1,657,600

2,176,400

3,745,500

5,749,100

30

70

200

700

2,000

7,000

20,000

70,000

30,000

0.0014

0.0019

0.0022

0.0027

0.0030

0.0042

0.0092

0.0187

0.0052

0.0485

ASM_Jawad_Ch06.indd

164

1,071

1,094

1,101

1,115

1,126

39,006

39,591

39,770

40,126

40,407

MTC

02/24/2009 04:22AM

From Table 6.5, the sum of (D/Td)k for the 130,000-hour cycle duration is 0.0485. Referring to Fig.

6.6 with (nc/Nd)j = 0.0 and (D/Td)k = 0.0485, it is seen that the expected life of 130,000 hours is well

within the acceptable limits.

It is of interest to compare the above results with the evaluation of the creep damage using the hours

for one cycle, 13,000, as a basis. In this case, the value of (D/Td) from Table 6.5 is equal to 0.0168.

For ten cycles, the value is 0.168, over a factor of 3 greater but still within acceptable limits.

Example 6.2

A 12-in. diameter high-pressure superheater header in a heat recovery steam generator is constructed of modied 9Cr steel (SA 335-P91) and built in accordance with ASME- I. The following

design and operating data is given:

Data

Longitudinal ligament efciency (for circumferential stress): Eo = 0.60.

Circumferential ligament efciency (for longitudinal stress): Eo = 0.85, y = 0.40.

Thickness, t = 1.125 in. (12-in. Sch-140 pipe).

Maximum stress concentration factor due to holes in the shell is taken as Ksc = 3.3. This value is

taken from the appendix on stress indices for nozzles in VIII-2.

Expected cycles = 10,000 (one full cycle per day for 25 years).

Expected life = 200,000 hours 25 years.

The design longitudinal stress in the header due to tube weight, liquid, and bending is assumed

as 7000 psi.

Assume, for simplicity, that the design and operating conditions are the same.

Assume that the temperature drops to 700F before the start of a new cycle.

Isochronous curves are given in Fig. 6.7 for 1000F.

The design fatigue strain range for 9Cr-1M0-V steel is given in Table 6.6.

The minimum stress-to-rupture values for 9Cr-1Mo-V steel are given in Table 6.7.

Determine whether the above conditions are adequate in accordance with the creep design rules in

Chapters 4, 5, and 6.

Solution

The various allowable stress values are:

S = 16,300 psi (II-D)

So = 16,300 psi (III-NH)

Sm = 19,000 psi (III-NH)

Smt = 14,300 psi for 200,000 hours (III-NH)

St = 14,300 psi for 200,000 hours (III-NH)

Sy = 53,200 psi at 700F (II-D), which is the cold end of a cycle

Sy = 40,200 psi at 1000F (II-D)

E = 27,500,000 psi at 700F (II-D)

E = 25,400,000 psi at 1000F (II-D)

(a) Check design thickness in accordance with ASME-I

From Eq. (4.1),

t

13001275

00

216300060 2041300

080 in. 1125 in.

ASM_Jawad_Ch06.indd

165

MTC

02/24/2009 04:22AM

166

Chapter 6

FIG. 6.7

ISOCHRONOUS STRESS-STRAIN CURVES FOR 9Cr-1Mo-V STEEL

AT 1000F (ASME, III-NH)

(b) Check load-controlled stress limits in accordance with III-NH (Fig. 4.8)

= Ro/Ri = 6.375/5.25 = 1.2143

Circumferential stress at inside surface with a ligament efciency of 0.60

membrane = Pm = P/(0.6)( - 1) = 10,110 psi

bending = Pb = 0

secondary = Q = 645/0.6 = 1075 psi

Longitudinal stress at inside surface with a ligament efciency of 0.85

membrane = Pm = P/(0.85)( 2 - 1) + 7000 = 10,225 psi

bending = Pb = 0

secondary = Q = 0

Radial stress at inside surface,

membrane = Pm = -P/2 = -650 psi

bending = Pb = 0

secondary = Q = -650 psi

The value of Pm is obtained from Eq. (4.14) as

Pm = 10,815

ASM_Jawad_Ch06.indd

166

PL + Pb = 10,815

and

MTC

02/24/2009 04:22AM

TABLE 6.6

DESIGN FATIGUE STRAIN RANGE FOR 9Cr-1Mo-V STEEL (ASME, III-NH)

Strain range, t [in./in. (m/m)] at 1000F

(540C)

Number of Cycles,1 Nd

10

20

40

102

2 102

4 102

103

2 103

4 103

104

2 104

4 104

105

2 105

4 105

106

2 106

4 106

107

2 107

4 107

108

1

0.028

0.019

0.0138

0.0095

0.0075

0.0062

0.0050

0.0044

0.0039

0.0029

0.0024

0.0021

0.0019

0.00176

0.0017

0.00163

0.00155

0.00148

0.00140

0.00132

0.00125

0.00120

Design limits

Pm < So

PL + Pb < 1.5So

10,815 psi < 24,450 psi

Operating limits

Pm < Smt

PL + Pb < 1.5Sm

PL + Pb/kt < St

10,815 psi < 28,500 psi

10,815 psi < 14,300 psi

TABLE 6.7

MINIMUM STRESS-TO-RUPTURE VALUES FOR 9Cr-1Mo-V STEEL (ASME, III-NH)

U.S. customary units

Temp., F 10 hr 30 hr 10 hr 3 102 hr 103 hr 3 103 hr 104 hr 3 104 hr 105 hr 3 105 hr

2

700

750

800

850

900

950

1000

1050

1100

1150

1200

71.0

69.0

66.5

63.4

59.8

51.2

42.8

35.6

29.2

23.7

19.0

ASM_Jawad_Ch06.indd

71.0

69.0

66.5

63.4

57.0

47.9

39.9

32.9

26.8

21.6

17.1

71.0

69.0

66.5

63.4

53.3

44.5

36.8

30.1

24.4

19.4

15.2

167

71.0

69.0

66.5

59.7

50.0

41.5

34.1

27.7

22.3

17.6

13.6

71.0

69.0

66.5

56.0

46.6

38.5

31.4

25.3

20.1

15.7

11.9

71.0

69.0

63.1

52.7

43.7

35.8

29.0

23.2

18.3

14.1

10.5

71.0

69.0

59.4

49.3

40.6

33.1

26.6

21.1

16.4

12.4

8.0

MTC

71.0

67.3

56.1

46.3

37.9

30.7

24.5

19.2

14.8

10.2

6.5

71.0

63.5

52.7

43.3

35.2

28.2

22.3

17.3

13.1

8.2

4.9

71.0

60.2

49.6

40.6

32.8

26.1

20.5

15.7

11.7

6.7

3.7

02/24/2009 04:22AM

168

Chapter 6

(c) Check strain- and deformation-controlled limits in accordance with III-NH (Fig. 4.8)

Average yield stress = 46,700 psi

Calculate X and Y from Eqs. (5.1) and (5.2)

The value of Q is obtained from Eq. (4.14) as

Q

1510 psi

10,815/46,700 = 0.23

1510 46700

003

X + Y = 0.26 < 1.0

There is no need to check the simplied inelastic analysis Tests B because the requirement of Test

A-2 is met.

Thus, the requirements of strain and deformation limits are met.

(d) Check creep-fatigue requirements using elastic analysis in accordance with III-NH, Eq. (6.11)

Calculate the value of 3Sm, which is the smaller of

3Sm = (3)(19,000) = 57,000 psi

or

3Sm

15Sm 05St

35650 psi

3Sm E

35600 25400000

00014

The requirements of Test A-2 are met.

The requirement of (PL + Pb + Q) 3Sm must be met.

The value of (PL + Pb + Q) is obtained from Eq. (4.14) as

PL Pb Q

Step 1

Total amount of hours, TH = 200,000 hours

Step 2

Hold temperature, T HT = 1000F

Step 3

Average cycle time = 24 hours/cycle.

Step 4

The maximum strain during the cycle is given by

Dmax = 2Salt/E = 12,030/25,400,000 = 0.000474 in./in.

Because the magnitude of Dmax is within the elastic limit, the value of Dmod can be written as

Dmod = KscDmax = (3.3)(0.000474) = 0.0016 in./in.

ASM_Jawad_Ch06.indd

168

MTC

02/24/2009 04:22AM

Step 5

From Fig. 6.2, let f be conservatively equal to 1.0.

From Fig. 6.3, Kv = 1.0.

Equation (6.9) gives Kv = 1.0.

From Eq. (5.11) with Z = X,

Vc

02353200 12240 psi

125 Vc

15300 psi

Enter the isochronous chart (Fig. 6.7) with a stress of 15,300 psi, and obtain the strain for either (a)

24 hours (duration of one cycle from step 3), which gives Dc = 0.00060, or (b) 200,000 hours, which

gives Dc = 0.0025 for the full life. The value for one cycle is then given by

Dc = 0.0025/10,000 0

t = (1.0)(0.0016) + (3.3)(0) = 0.0016

Step 6

In Fig. 6.7, stress-versus-time values are obtained as shown in Table 6.8.

Step 7

Equation (6.11) is solved in this step. The rst part of the equation, (nc/Nd)j, is rst obtained as follows:

First part of Eq. (6.11)

nc = 10,000 cycles

The value of Nd is obtained from Table 6.6 with t = 0.0016. A cycle life of 1,000,000 is obtained

at a temperature of 1000F.

(nc/Nd)j = 10,000/1,000,000 = 0.01

TABLE 6.8

STRESS-VERSUS-TIME VALUES

ASM_Jawad_Ch06.indd

Time (hr)

Stress (ksi)

Sj

1

3

10

30

100

300

1,000

3,000

10,000

30,000

100,000

200,000

36.5

30.0

28.0

26.7

25.3

23.3

22.0

21.0

19.3

18.0

16.0 (use 16.5)

14.7 (use 16.5)

13.7 (use 16.5)

169

MTC

02/24/2009 04:22AM

170

Chapter 6

The second part of Eq. (6.11), (D/Td)k, is determined numerically from Table 6.8 and from the

differential equation dT/Td. Calculations for 200,000 hours are summarized in Table 6.9.

Column A in Table 6.9 gives the incremental locations for calculating stresses from Table 6.8.

Column B shows the stress values obtained from Table 6.8 and column H lists the incremental duration assumed for these stresses. Column C adjusts the stress values by the K factor and column G

lists the hours obtained from Table 6.7 using the stress values in column C. Column I is the ratio of

column H over column G, which is the numerical value of the quantity (dT/Td).

The 1000F line in Table 6.7 terminates at the 20,500 psi stress level. Stress levels in column C

drop below the 20,500 psi level in rows 1 through 12. The Larson-Miller parameter, PLM, is used

to obtain approximate rupture life by using higher temperature levels given in Table 6.7. These

calculations are shown in columns D, E, and F. Column D lists the temperatures from Table 6.7

corresponding to the stress levels in column C. Column E calculates the corresponding PLM from

Eq. (1.3) for the temperature shown in column D at 300,000 hours. Column F recalculates the hours

corresponding to a temperature of 1000F from Eq. (1.3) with the corresponding PLM factor.

The quantity (3Sm)/E, calculated earlier, is equal to 0.0014, whereas t is equal to 0.0016. Because

t > (3Sm)/E, shakedown does not occur and the peak stress does reset on each cycle. Under this

scenario, peak stress can be assumed to relax through each cycle and the cumulated creep damage

is given by Table 6.9.

In Table 6.9, the value of (D/Td)k for 24 hours is equal to 0.059. Note that most of the computed

creep damage occurs in the rst hour. This computed value can be more realistically evaluated by taking smaller time steps and computing the damage at each step. In this case, time steps of 0.25 hours

will reduce the damage during the rst hour by more than a factor of two. However, even with the rst

hour damage reduction the total value of (D/Td)k for 10,000 cycles is approximately 350, which is

inadequate. Thus, the required thickness needs to be increased to accommodate 200,000 hours.

Second iteration

A new trial thickness may be assumed by multiplying the original thickness by the ratio of strain t over

t = (1.125)(0.0016/0.0014) = 1.286 in.

TABLE 6.9

NUMERICAL CALCULATIONS OF (T/Td)k

A

Location

(hr)

Sr (ksi)

Sr/0.9

(ksi)

Equivalent

temp. (F)

PLM

Time (hr)

Table 6.7

(hr)

Table 6.8

(hr)

H/G

36.5

30.0

28.0

26.7

25.3

23.3

22.0

21.0

19.3

18.0

16.5

16.5

16.5

40.5

33.3

31.1

29.7

28.1

25.9

24.4

23.3

21.4

20.0

18.3

18.3

18.3

366,750

756,400

756,400

23

507

1,250

2,417

5,625

16,670

33,180

68,180

200,000

366,750

756,400

756,400

1

2

7

20

70

200

700

2,000

7,000

20,000

70,000

100,000

0.043

0.004

0.006

0.008

0.012

0.012

0.021

0.029

0.035

0.055

0.093

0.132

Sj

1

3

10

30

100

300

1,000

3,000

10,000

30,000

100,000

200,000

Total

ASM_Jawad_Ch06.indd

1,005

1,023

1,023

37,324

37,783

37,783

0.450

170

MTC

02/24/2009 04:22AM

Try a 12-in. Sch-160 pipe with OD = 12.75 in. and t = 1.3125 in.

(a) Check design thickness in accordance with ASME-I

By inspection, the newer thickness is adequate as per ASME-I.

(b) Check load-controlled stress limits in accordance with III-NH (Fig. 4.8)

= Ro/Ri = 6.375/5.0625 = 1.2593

Circumferential stress at inside surface with a ligament efciency of 0.60

membrane = Pm = P/(0.6)( - 1) = 8355 psi

bending = Pb = 0

secondary = Q = 645/0.6 = 1075 psi

Longitudinal stress at inside surface with a ligament efciency of 0.85

membrane = Pm = P/(0.85)( 2 - 1) + 7000 = 9610 psi

bending = Pb = 0

secondary = Q = 0

Radial stress at inside surface

membrane = Pm = -P/2 = -650 psi

bending = Pb = 0

secondary = Q = -650 psi

The value of Pm is obtained from Eq. (4.14) as

Pm = 9690 psi,

Design limits

Pm < So

PL + Pb < 1.5So

9690 psi < 24,450 psi

Operating limits

Pm < Smt

PL + Pb < 1.5Sm

PL + Pb/kt < St

9690 psi < 28,500 psi

9690 psi < 14,300 psi

and PL + Pb = 9690

(c) Check strain- and deformation-controlled limits in accordance with III-NH (Fig. 4.8)

Average yield stress = 46,700 psi

Calculate X and Y from Eqs. (5.1) and (5.2).

The value of Q is obtained from Eq. (4.14) as

Q = 1510 psi,

X = 10,815/46,700 = 0.23

X + Y = 0.24 < 1.0

There is no need to check the simplied inelastic analysis Tests B because Test A-2 is met.

Thus, the requirements of strain and deformation limits are met.

(d) Check creep-fatigue requirements using elastic analysis in accordance with III-NH, Eq. (6.11)

Calculate the value of 3Sm, which is the smaller of

ASM_Jawad_Ch06.indd

171

MTC

02/24/2009 04:22AM

172

Chapter 6

3Sm = (3)(19,000) = 57,000 psi

or

_

3Sm = 1.5Sm + 0.5St = 35,650 psi

_

3S m /E = 35,600/25,400,000 = 0.0014

The requirements of Test A-2 are met.

The requirement of (PL + Pb + Q) 3Sm must be met.

The value of (PL + Pb + Q) is obtained from Eq. (4.14) as

(PL + Pb + Q) = 10,820 psi,

Step 1

Total amount of hours, TH = 200,000 hours

Step 2

Hold temperature, T HT = 1000F

Step 3

Average cycle time = 24 hours/cycle

Step 4

The maximum strain during the cycle is given by

Dmax = 2Salt/E = 10,820/25,400,000 = 0.000426 in./in.

Because the magnitude of Dmax is within the elastic limit, the value of Dmod can be written as

Dmod = Ksc Dmax = (3.3)(0.000426) = 0.0014 in./in.

Step 5

From Fig. 6.2, let f be conservatively equal to 1.0.

From Fig. 6.3, Kv = 1.0.

Equation (6.9) gives Kv = 1.0.

From Eq. (5.11) with Z = X,

Vc

02153200

1.25 Vc

11200 psi

14000 psi

Enter the isochronous chart (Fig. 6.7) with a stress of 14,300 psi, and obtain the strain for either (a)

24 hours (duration of one cycle from step 3), which gives Dc = 0.00056, or (b) 200,000 hours, which

gives Dc = 0.0015 for the full life. The value for one cycle is then given by

Dc = 0.0015/10,000 0

t = (1.0)(0.0014) + (3.3)(0) = 0.0014

ASM_Jawad_Ch06.indd

172

MTC

02/24/2009 04:22AM

TABLE 6.10

STRESS-VERSUS-TIME VALUES

Time (hr)

Stress (ksi)

Sj

1

3

10

30

100

300

1,000

3,000

10,000

30,000

100,000

200,000

34.0

28.7

26.7

25.3

24.0

22.7

21.3

20.0

18.7

17.3

16.0 (use 16.5)

14.0 (use 16.5)

13.7 (use 16.5)

Sj = 34.0 ksi.

Step 6

In Fig. 6.7, stress-versus-time values are obtained as shown in Table 6.10.

Step 7

Equation (6.11) is solved in this step. The rst part of the equation, (nc/Nd)j, is obtained rst as

follows:

First part of Eq. (6.11)

nc = 10,000 cycles

The value of Nd is obtained from Table 6.6 with t = 0.0014. A cycle life of 10,000,000 is obtained

at 1000F.

(nc/Nd)j = 10,000/10,000,000 = 0.001 0

The second part of Eq. (6.11), (D/Td)k, is determined numerically from Table 6.10 and from the

differential equation dT/Td. Calculations for 200,000 hours are summarized in Table 6.11.

The quantity (3Sm)/E, calculated earlier, is equal to 0.0014 and t is also equal to 0.0014. Because

t = (3Sm)/E, shakedown does occur and the peak stress does not reset on each cycle. Under this

scenario, peak stress can be assumed to relax monotonically throughout the entire design life and

the cumulated creep damage is given by Table 6.11.

The total value of (D/Td)k from Table 6.11 is 0.358. This value is acceptable from Fig. 6.6 for an

expected life of 200,000 hours.

The result indicates that the required original thickness of 1.125 in. is inadequate for a 10,000

cycle service with 200,000 hours at 1000F. The new thickness of 1.3125 in. is adequate for the

intended service.

ASM_Jawad_Ch06.indd

173

MTC

02/24/2009 04:22AM

174

Chapter 6

TABLE 6.11

NUMERICAL CALCULATIONS OF (T/Td)k

A

Location

(hr)

Sj

1

3

10

30

100

300

1,000

3,000

10,000

30,000

100,000

200,000

Total

6.3

Sr (ksi)

Sr/0.9

(ksi)

Equivalent

temp. ()

34.0

28.7

26.7

25.3

24.0

22.7

21.3

20.0

18.7

17.3

16.5

16.5

16.5

37.8

31.9

29.7

28.1

26.7

25.2

23.7

22.2

20.8

19.2

18.3

18.3

18.3

1,014

1,023

1,023

E

PLM

37,553

37,783

37,783

Time (hr)

Table 6.7

(hr)

Table 6.8

(hr)

H/G

526,530

756,400

756,400

77

870

2,417

5,625

9,708

23,330

55,460

111,110

266,670

526,530

756,400

756,400

1

2

7

20

70

200

700

2,000

7,000

20,000

70,000

100,000

0.013

0.002

0.003

0.004

0.007

0.009

0.013

0.018

0.026

0.038

0.093

0.132

0.358

WELDED COMPONENTS

The procedure outlined in Section 6.2 is also applicable to welded components. Welded reduction

factor, Rw, must be incorporated into stress calculations for creep rupture.

6.4

In Section 6.2 the analysis of components subjected to a repetitive cyclic load was presented. Repetitive cyclic in that context referred to a number of cycles of the same type and magnitude, such

as startup followed by a period of operation and then shutdown. The remainder of this chapter will

address the situation where there are two or more types of cycles with varying magnitudes for example, normal startup and shutdown cycles of a xed number and magnitude, and a different number

of thermal transient cycles of a different number and magnitude.

There are a number of methods developed to combine cyclic histories, many of which are described

in some detail in Bannantine (1990). The essential feature of these methods is to account for the additive effect of combining strain ranges. As discussed in the ASME criteria for Division 2 (1969), When

stress cycles of various frequencies are intermixed through the life of a vessel it is important to identify

correctly the number and range of each type of cycle. It must be remembered that a small increase in

stress range can produce a large decrease in fatigue life, and this relationship varies for different portions of the fatigue curve. Therefore the effect of superposing two stress amplitudes cannot be evaluated by adding the usage factors obtained from each amplitude by itself. The stresses must be added

before calculating the usage factors. Although the above is written in terms of stress, the discussion

is equally applicable in terms of strain frequency and range.

Example 6.3

Consider the case of a thermal transient occurring in a pressurized vessel. At a point in the vessel,

the peak stress due to pressure is 20,000 psi tension and the added stress from the thermal transient is

70,000 psi tension. The thermal stress occurs 10,000 times and the pressure stress occurs 1000 times.

What is the combined effect?

ASM_Jawad_Ch06.indd

174

MTC

02/24/2009 04:22AM

Solution

Add the thermal stress range to the pressure stress range for 1000 cycles. Thus, the total usage factor

is the sum of the usage factor for 1000 cycles with a stress range of 90,000 psi and the usage factor for

9000 cycles with a stress range of 70,000 psi.

6.5

There are several locations in the ASME Code where the issue of combining stresses with different

frequencies and ranges are discussed. In Subsection NH, procedures are given in T-1413 and T-1414

for determining equivalent strain range; T-1413 for the more general case where principal strains

change direction, and T-1414 for the simpler case where they do not. These procedures are applicable

for both elastic and inelastic methodologies. Section III-NH, T-1432 also permits the use of the stress

difference procedure in III-NB-3216 when evaluating creep-fatigue limits using elastic analysis. Interestingly, Section VIII Div 2 also identies additional procedures for combining cycle types in Annex

5.B and in Annex 5.C. Although some of these procedures read quite differently from others, they

all accomplish the objective described in Section 6.1; they provide methodologies for superimposing

strain amplitudes to ensure that the usage factor accounts for reduction in fatigue life due to additive

stress or strain cycling. Because the evaluations discussed herein are based on elastic analysis and the

calculation results expressed in terms of stress, the procedures of NB-3216 are the procedures that will

form the basis for the following discussion.

6.6

When the design specication delineates a specic loading sequence, then such sequence should

be used to determine the equivalent strain ranges. If the sequence is not specied, then the following

procedures should be used.

6.6.1

Step 1

Calculate all strain components for each point, I, in time (xi, yi, zi, xyi, yzi, zxi) for the complete

cycle. Note that when conducting elastic analysis, which is the basis of this discussion, peak strains

from geometric discontinuities are not included because these effects are accounted for later in the

procedure.

Step 2

Select a point when conditions are at an extreme for the cycle, either maximum or minimum, and

refer to this time point by the subscript o.

Step 3

Calculate the history of the change in strain components by subtracting the values at time o from

the corresponding components at each point in time, i, during the cycle.

Dxi = xi - xo

Dyi = yi - yo

Dzi = zi - zo,

etc.

ASM_Jawad_Ch06.indd

175

MTC

02/24/2009 04:22AM

176

Chapter 6

Step 4

Calculate the equivalent strain range for each point in time as

2

2

2

+ Dyzi

Dzxi

)]

Dequiv,i = [0.707/(1 + *)][(Dxi - Dyi)2 + (Dyi - Dzi)2 + (Dzi - Dxi)2 + 1.5(Dxyi

(6.12)

Not Rotate

Step 1

Determine the principal strains versus time for the cycle.

Step 2

At each time interval of step 1, determine the strain differences (1 - 2), (2 - 3), (3 - 1).

Step 3

Select a point when conditions are at an extreme for the cycle, either maximum or minimum, and

refer to this time by the subscript o.

Step 4

Calculate the history of the change in strain differences by subtracting the values at time o from the

corresponding components at each point in time, i, during the cycle. Designate these strain difference

changes as

D(1 - 2)i = (1 - 2)i - (1 - 2)o

D(2 - 3)i = (2 - 3)i - (2 - 3)o

D(3 - 1)i = (3 - 1)i - (3 - 1)o

Step 5

For each point in time i, calculate the equivalent strain range as

Dequiv, i = [0.707/(1 + *)]{[D(1 - 2)i]2 + [D(2 - 3)i]2 + [D(D3 - 1)i]2}1/2

(6.13)

Elastic Analysis

6.6.3.1

Step 1

Determine the principal stresses versus time for the cycle.

Step 2

At each time interval of step 2, determine the stress differences (1 - 2), (2 - 3), (3 - 1).

Step 3

Calculate the history of the change in stress differences by subtracting the values at a reference time

from the corresponding components at each point in time, i, during the cycle. The maximum stress

ASM_Jawad_Ch06.indd

176

MTC

02/24/2009 04:22AM

value of these differences is obtained from Eq. (4.14) and is equal to the value of 2Salt, which is the

starting point of Step 4 under Section 6.2 CREEP-FATIGUE EVALUATION USING ELASTIC

ANALYSIS.

6.6.3.2 Rotating Principal Stress Direction. This procedure is basically the same as the procedure for rotating principal strain direction. The stress differences are rst determined at the component level for all six components and then the principal stresses are determined from those component

stress differences.

6.6.3.3 Variable Cycles. The following is based on the guidance cited in III-NH by reference to

III-NB-3222.4 Analysis for Cyclic Operation.

If there are two or more types of stress cycles which produce signicant stresses, their cumulative

effect shall be evaluated as shown:

Step 1

Designate the specied number of times each type of cycle (types 1, 2, 3, , n) will be repeated during the life of the component as n1, n2, n3, , nn, respectively.

Note: In determining n1, n2, n3, , nn, consideration shall be given to the superposition of cycles of

various origins that produce a total stress difference range greater than the stress difference range

of the individual cycles.

Step 2

For each type of stress cycle, determine the alternating stress intensity, Salt, and designate these

quantities Salt 1, Salt 2, Salt 3, , Salt n.

Step 3

Proceed with evaluation as per Step 4 under Section 6.2 CREEP-FATIGUE EVALUATION

USING ELASTIC ANALYSIS.

Example 6.4

Using the data previously developed for Examples 5.1, 5.2, and 6.1 to evaluate the life cycle of the

2.0-in. thick shell for 130,000 hours when subjected to the original maintenance shutdown cycle plus

the unplanned process-induced shutdown as described below.

Pressure. The pressure remains constant.

Temperature. The temperature decreases such that the inner wall is 36F colder than the outer wall

as shown in Fig. 6.8. The process is then re-established and operation returned to normal conditions.

This cycle, cool down and restart, takes approximately 8 hours to complete. This cycle is expected to

occur, on average, six times during each 18-month operational cycle.

Solution

The stress components for Example 5.1 for the outside surface of the 2.0-in. shell are tabulated in

Table 6.12 with an additional column for the 36F DT condition.

Step 1

Principal stresses for each time point in the cycle as identied in Fig. 6.8 are computed from component stresses in Table 6.12 as shown below at time 2.

ASM_Jawad_Ch06.indd

177

MTC

02/24/2009 04:22AM

178

Chapter 6

FIG. 6.8

PRESSURE AND TEMPERATURE HISTORY

= P m

L = P mL + QbL = 2250 + 14,975 = 16,850

R = P mr + Qbr = (-150) + 150 = 0

Step 2

Principal stress differences, or stress intensities, are computed from the principal stress as shown

below for time 2.

- L = 19,475 - 16,850 = 2250

L - R = 16,850 - 0 = 16,850

R - = 0 - 19,475 = -19,475

ASM_Jawad_Ch06.indd

178

MTC

02/24/2009 04:22AM

TABLE 6.12

STRESS OUTSIDE OF THE 2.0-in. SHELL

Stress

Membrane hoop stress, Pm

Membrane axial stress, PmL

Membrane radial stress, Pmr

Circumferential bending stress, Qb

Axial bending stress, QbL

Radial bending stress, Qbr

Pressure

stress

Thermal stress,

T = 90F

Thermal stress,

T = 36F

4,500

2,250

-150

0

0

0

14,975

14,975

0

0

0

0

-5,990

-5,990

0

0

150

Step 3

Determine the range of principal stress differences between the reference points in time dening

the cycles of interest. Shown below is the evaluation of the cycle range dened by time points 1 and

2. Note that this corresponds to the cycle range for Example 6.1. Because stress values at time point 1

are zero, the absolute value of the difference is straightforward.

( - L)2 - ( - L)1 = 2250 - 0 = 2250

(L - R)2 - (L - R)1 = 16,825 - 0 = 16,825

(R - )2 - (R - )1 = -19,475 - 0 = 19,475

From Eq. (4.14), the effective stress is 2Salt 1-2 = 16,060 psi.

Stress differences (Table 6.13) are identied for the cycle dened by time points 1 and 2 in the

fourth column (Sr1-2); for the cycle dened by time points 3 and 4 in column 7 (Sr3-4); and for the composite cycle dened by time points 1, 2, 3, and 4 in the last column (Sr1-4).

From the above table, the maximum absolute values of the cycle range = 2Salt are obtained from Von

Mises equation (Eq. 4.14):

2Salt 1-2 = 16,060 psi

2Salt 3-4 = 1490 psi

2Salt 1-4 = 16,260 psi

Cycle 1-4 with a cycle time of 13,000 hours and cycle 3-4 with a negligible cycle time will be selected to represent the combined loading history. The selection of combined cycles requires consideration of the cycle time and thus goes beyond the normal superposition considerations associated with

nontime-dependent design criteria. This point will be discussed further when considering the evaluation of creep damage.

The following evaluation follows the same procedure used to evaluate creep-fatigue damage in Example 6.1.

Step 4

First, consider cycle dened by 2Salt 1-4 = 16,260 psi

TABLE 6.13

PRINCIPAL STRESSES AND ALTERNATING STRESS INTENSITY

Time 1

Time 2

Sr1-2

Time 3

Time 4

Sr3-4

Sr1-4

0

0

0

0

0

0

19,475

16,850

0

2,250

16,825

-19,475

2,250

16,825

19,475

4,500

2,250

0

2,250

2,250

-4,500

-1,490

-3,740

0

-2,250

-3,740

1,490

4,500

5,990

5,990

4,500

20,565

20,965

L

R

- L

L - R

R -

ASM_Jawad_Ch06.indd

179

MTC

02/24/2009 04:22AM

180

Chapter 6

Dmax = 2Salt 1-4 /E = 16,260/22,400,000 = 0.000726

Next, evaluate

2

Dmax = (S */S)K sc

Dmax (Method 1)

However, as before, Dmax and KscDmax are within the proportional limit of the composite stressstrain curve, and Dmod reduces to

' H mod

110000726

000080

Step 5

This step is essentially identical to the steps used in Example 6.1, and the assumptions therein are

equally applicable. For Dmod = 0.00080, the resulting expression for t becomes:

t = (1.0)(0.00080) + (1.1)(0.000026) = 0.000829

Note that the creep strain increment per cycle, Dc, remains the same because the pressure stress is

the same and the value of Z in region E is equal to X.

From Fig. 5.6, the time independent stress Sj 15,000 psi.

Before proceeding with the creep damage assessment, it is necessary to assess the value of t and Sj

for the remaining cycle.

Step 3

From the preceding step 3, the stress intensity range for cycle 3-4 is given by

2Salt 3-4 = 1490 psi

Step 4

As before,

Dmax = 2Salt 3-4 /E = 1490/22,400,000 = 0.0000665

' H mod

1100000665

0000073

Step 5

As before, for Dmod = 0.000073

t = (1.0)(0.000073) + (1.1)(0.000026) = 0.00010

Note that, per the presumed composite cycle denition, there is no signicant cycle time and the

incremental creep strain Dc will be equal to zero.

Conservatively based on t 3-4 = 0.00010

Sj = 2240 psi

Step 6

It is clear that cycle 3-4 satises the shakedown criteria and the assumed superposition is valid. Noting that t 1-4 = 0.000829 also satises the shakedown criteria, the resulting stress history for evaluating

ASM_Jawad_Ch06.indd

180

MTC

02/24/2009 04:22AM

creep damage is quite closely approximated by the values in Table 6.5, and the overall creep damage

fraction will be:

(D/Td)k = 0.0485

In summary, the unplanned process-induced shutdowns have a very slight impact on the cyclic

stress range that is wiped out within the rst hour of sustained operation. The impact on vessel life

is negligible. Physically, the loading for Example 6.1 resembles a hold time creep-fatigue test where

there is no subsequent yielding after the initial startup sequence. The additional cycles in this example

resemble additional minor, and more frequent, dips in the strain level in the creep-fatigue test with no

apparent impact on cyclic life.

Problems

6.1 The hydrotreater given in Problem 5.1 is shut down after every 3 years (25,000 hours) of operation for maintenance. The shutdown period is 2 weeks. Accordingly, the expected number

of cycles is 8. Evaluate the life cycle of 200,000 hours at 950F.

6.2 The boiler header given in Problem 5.2 is shut down after every 2.5 years (20,000 hours) of

operation for maintenance. The shutdown period is 6 weeks. Accordingly, the expected number of cycles is 15. Evaluate the life cycle of 300,000 hours at 975F.

ASM_Jawad_Ch06.indd

181

MTC

02/24/2009 04:22AM

ST. LOUIS, MISSOURI)

ASM_Jawad_Ch07.indd

182

MTC

02/24/2009 05:17AM

CHAPTER

7

MEMBERS IN COMPRESSION

7.1

INTRODUCTION

The analysis of column buckling in the creep range is an extremely complicated subject. Different

methods of analysis have been proposed and are available to the designer. Each of these methods

has its advantages and disadvantages regarding accuracy of results and complexity of solution. The

methods presented in this chapter are relatively simple to perform but fairly conservative. Designers

must exercise their judgment and use their experience in using the various equations presented in this

chapter.

7.2

DESIGN OF COLUMNS

Background equations for the design of axially loaded members operating at temperatures below

the creep range are presented in Section 7.2.1. These equations are then used as basis for developing

buckling equations for axially loaded members operating at temperatures above the creep range as

discussed in Section 7.2.2. The creep buckling phenomenon of columns can be thought of in terms of

the column creeping under a sustained load up to a deformation level where regular buckling occurs.

Accordingly, a time factor must be included in the buckling equations as discussed below.

7.2.1

The elastic buckling equation for members loaded axially (Fig. 7.1) at temperatures below the creep

range is given by

d2y/dx2 = -M/EI = -Py/EI

(7.1)

(d2y/dx2) + k2y = 0

(7.2)

where

E

I

M

P

x

y

= modulus of elasticity

= moment of inertia of member

= moment in column

= applied load

= length

= deection

or

where

k2 = P/EI

183

ASM_Jawad_Ch07.indd

183

MTC

02/24/2009 05:17AM

184

Chapter 7

FIG. 7.1

SIMPLY SUPPORTED COLUMN

y = A sin(kx) + B cos(kx)

(7.3)

Substituting the rst boundary condition of y = 0 at x = 0 into Eq. (7.3) gives B = 0. Substituting the

second boundary condition of y = 0 at x = L into Eq. (7.3) gives a non-trivial solution of the critical

axial load with the lowest value of

Pcr

S 2 EI

L2

(7.4)

Equation (7.4), referred to as Eulers equation, can be written in terms of effective length and critical stress as

V cr

S 2E

(7.5)

l r 2

where

A

l

r

cr

= area of member

= effective length of column

= radius of gyration = (I/A)0.5

= critical stress

The Steel Construction Manual of the American Institute of Steel Construction (AISC, 1991) uses

Eq. (7.5) as the basis for designing slender columns. Columns made of carbon steel with a yield stress

of 36 ksi are considered slender when (l/r) > 126. The actual design equation, with a factor of safety

of 1.92, is given by

sa =

12 p 2E

23(l /r )2

for (l /r ) > C c

(7.6)

where

Cc

2 S 2E

Sy

1 2

and

Sy = yield stress of the material

a = allowable compressive stress

A plot of Eq. (7.6) is shown in Fig. 7.2. The buckling strength tends to approach innity as the slenderness ratio (l/r) becomes smaller, as is the case with short columns. However, the allowable stress

ASM_Jawad_Ch07.indd

184

MTC

02/24/2009 05:17AM

FIG. 7.2

ALLOWABLE COMPRESSIVE STRESS FOR STEEL WITH SY = 36 KSI

for short columns is controlled by yield stress rather than buckling. Accordingly, the AISC limits the

allowable compressive stress for short columns to a maximum value of (Sy/1.67). A parabolic equation is then generated to connect the allowable compressive stress from a value of (Sy/1.67) for short

columns to the Euler equation given by Eq. (7.6) for slender columns. The point of tangency of the two

equations is at (l/r) = 126 for low strength steels as shown in Fig. 7.2. The actual parabolic equation

is given by AISC as

2

a

1 05C 1 Sy

3

for l r C c

(7.7)

where

C1

l r

Cc

Equations (7.6) and (7.7) can be approximated for design of non-pressure parts in pressure vessels.

A factor of safety of 2.0 can be used for Eulers equation (Eq. 7.6). Equation (7.7) can be simplied by

setting it equal to the allowable stress, which is determined from the tensile, yield, creep, and rupture

criteria. Hence, the two equations can be combined into a single equation

Sb =

p 2E

2(l /r )2

(7.8)

where

S = allowable tensile stress given in the American Society Mechanical Engineers (ASME) code

Sb = allowable compressive stress

Figure 7.3 shows a plot of Eq. (7.8).

Example 7.1

A pressure vessel is operating at 1200F has a 316 stainless steel internal support bracket as shown

in Fig. 7.4. Determine the size of member AB due to an applied load of 450 lb. E = 21,200 ksi, S =

7.4 ksi.

ASM_Jawad_Ch07.indd

185

MTC

02/24/2009 05:17AM

186

Chapter 7

FIG. 7.3

PLOT OF BUCKLING STRESS IN A SIMPLY SUPPORTED COLUMN

Solution

The load on member AB from Fig. 7.4 is 900 lb.

Minimum area required = 900/7400 = 0.12 in.2.

Try a -in. diameter bar.

A = R2 = 0.196 in.2

r = R/2 = 0.125 in.

= 4600 psi

From Eq. (7.8), the allowable compressive stress is

Sb

S 2 21200000

2200125 2

FIG. 7.4

INTERNAL BRACKET

ASM_Jawad_Ch07.indd

186

MTC

02/24/2009 05:17AM

A = R2 = 0.31 in.2

r = R/2 = 0.156 in.

= 2900 psi

Sb

S 2 21200000

2200156 2

7.2.2

The analysis of columns operating in the creep regime (Odqvist, 1966) is extremely complicated due

to numerous factors, some of which are not necessarily well known. These include column eccentricity, effect of primary and secondary creep on buckling, interaction between elastic and creep buckling,

and behavior of material properties with time at elevated temperatures. These factors require numerous

assumptions that would allow the designer to develop a design criterion for axial compression in the

creep range. Some of these factors and assumptions that are of particular interest to the designer are:

Creep buckling in columns with an eccentricity will occur at any axial compressive load

(Kraus, 1980) no matter how small it is when the column is subjected to temperatures in the

creep range over a certain period.

Creep buckling does not occur instantaneously (Boyle and Spence, 1983), but rather after a

certain time lapse. Thus, the design premise is based on determining a critical period that is

longer than the intended service of a component, rather than critical buckling force.

The Euler elastic buckling equation does not apply in the creep range due to the non-linear

relationship between stress and strain.

Creep buckling occurs at a specied time only for non-linear stress-strain behavior.

The tangent modulus, Et, gives a simplied equation for designing in the creep regime.

The shape of the deformed column is assumed (Flugge, 1967) to be sinusoidal.

Based on these facts, simplied equations for column buckling can be derived and adopted for

design purposes. One such method, developed by Kraus (1980), is discussed in this section. Detailed

evaluation of Krauss method by Dr. Don Grifn of Westinghouse showed that the results obtained

theoretically may differ from the results obtained from test data unless careful consideration is given

to various parameters effecting the buckling.

The general differential equation for bending due to axial compression is given by

d2y/dx2 = -M/EI = -P(yo + y)/EtI

(7.9)

where

Et

I

M

P

x

y

yo

= moment of inertia of member

= moment in column

= applied load

= length

= deection

= initial deection

Let the solution for y and yo be represented by the following Fourier series

yo = Am sin(mx/L)

ASM_Jawad_Ch07.indd

187

MTC

02/24/2009 05:17AM

188

Chapter 7

and

y = Bm sin(mx/L)

These two expressions satisfy the boundary conditions of a simply supported column. Substituting

these two expressions into Eq. (7.9) gives

Bm Am Bm

PL2

m2 S 2 E

tI

sinmS xL

This equation is satised if the rst bracketed term is set to zero. This gives

Bm

Am

m2 S 2 E

tI

PL2

(7.10)

1

The value of Bm in Eq. (7.10) approaches innity as the denominator approaches zero. Thus,

Pcr

m 2 S 2 Et I

(7.11)

L2

S 2 Et I

Pcr

L2

(7.12)

This equation is similar to the elastic buckling expression given by Eq. (7.4) with the exception of

the term Et. Equation (7.12) can be written in terms of critical stress, cr, as

V cr

S 2 Et

l r 2

(7.13)

where

A = area of member

l = effective length

r = radius of gyration = (I/A)0.5

The value of Et in Eq. (7.13) can be obtained from isochronous curves by using Nortons equation

for creep given by

= kn

(7.14)

where

k

n

= constant

= creep exponent, which is a function of material property and temperature

= strain rate (=d/dT )

= stress

- o = k nT

(7.15)

where

T = time

= strain

o = initial strain at T = 0

ASM_Jawad_Ch07.indd

188

MTC

02/24/2009 05:17AM

Et = d/d

(7.16)

d = knn-1Td

(7.17)

1

Et

k n V n 1 T

(7.18)

Combining this equation with Eq. (7.13) for creep buckling gives

S2

n

V cr

k nT l r 2

(7.19)

The constant k in Eq. (7.19) is obtained from Nortons equation (Eq. 7.14). However, for many

engineering problems a stationary stress condition exists as described in Chapters 1 and 2. Thus, for

such conditions the constant K can also be determined from the expression

= Kn

(7.20)

Constants K and n for any given material are easily obtained from isochronous curves similar to the

one shown in Fig. 7.5. This is accomplished by choosing two points on any particular curve and solving Eq. (7.20) for K and n. Once K and n are known, then Eq. (7.19) can be solved for the buckling

stress. Tests by Dr. Grifn of Westinghouse indicate that more accurate results are obtained when the

two points chosen on an isochronous curve are in the vicinity of the buckling strain. Otherwise, the

results tend to be on the unsafe side.

For design purposes, a factor of safety (FS) is added to Eq. (7.19) and it becomes

sb =

p2

(FS)K nT (l /r ) 2

1/ n

(7.21)

Example 7.2

Find the required diameter of member AB in Example 7.1 using the creep buckling Eq. (7.21) and

the isochronous curves in Fig. 7.5. Let S = 7400 psi, member force = 900 lb, expected life = 30,000

hours, and factor of safety = 1.5.

Solution

Two points, A and B, are chosen at random on the 30,000-hour curve in Fig. 7.5. The stress and

corresponding strain values at these two points are

A = 0.00013

at

A = 2700 psi

B = 0.00062

at

B = 4000 psi

Substituting these values into Eq. (7.20) and solving for n and K results in

n 4.0

and

K = 2.45 10-18

A = (2.0)2/4 = 3.14 in.2

Actual stress, = 900/3.14 = 290 psi

r = R/2 = 0.500 in.

l = 20.0 in.

ASM_Jawad_Ch07.indd

189

MTC

02/24/2009 05:17AM

190

Chapter 7

FIG. 7.5

ISOCHRONOUS STRESS-STRAIN CURVES FOR 316 STAINLESS STEEL AT

1200F (ASME, III-NH)

From Eq. (7.21), allowable stress is

Vb

S2

15 245 u

140

345 psi, which is greater than the actual stress of 290 psi

Note that this diameter is three times larger than that obtained in Example 7.1 for buckling below

the creep range.

ASM_Jawad_Ch07.indd

190

MTC

02/24/2009 05:17AM

COMPRESSION

The design equations presented in this method are based on ASME. The cylinder length is considered much larger than the radius. Thus, for axial compression the length does not appear in the equations. For short cylinders, the results obtained in this section are very conservative.

Creep Range

The ASME design equations for the axial buckling of cylindrical shells at temperatures below the

creep range are based on Sturms work (Sturm, 1941). Sturm reduced the differential equations for the

elastic axial buckling of a cylindrical shell with a large L/Do ratio to the form

crL = 0.63E(t/Ro)

(7.22)

063

Ro t

(7.23)

or

H cr

Vcr E

where

E = elastic modulus of elasticity

Ro = outside radius

t

= thickness of cylinder

crL = axial buckling stress

The ASME criteria for axial compressive loads at temperatures below the creep range use Eq. (7.23)

for elastic buckling. This is accomplished by splitting Eq. (7.22) into two parts and adding factors of

safety as follows

063

FS 1 Ro t

B = AE/(FS2)

(7.24)

(7.25)

where

A = allowable strain

B = allowable compressive stress

ASME uses a total factor of safety of 10 for axial compressive calculations with FS1 =5.0 and FS2 =

2.0. Thus, Eqs. (7.24) and (7.25) become

A

0125

Ro t

B = AE/2

(7.26)

(7.27)

Equation (7.27) is applicable only in the elastic region. ASME provides pseudo stress-strain curves

based on the tangent modulus in the plastic region. This is accomplished by using a stress-strain curve

for various given temperatures plotted in terms of elastic modulus, E, and tangent modulus, Et, past

the proportional limit. An example of such a chart for medium strength carbon steels is shown in Fig.

7.6. The ASME procedure for calculating allowable compressive stress is to determine rst strain A

from Eq. (7.26). Next, a value of allowable compressive stress B is determined from either Eq. (7.27)

for elastic buckling or from the tangent modulus lines of Fig. 7.6 when A falls past the material proportional limit.

ASM_Jawad_Ch07.indd

191

MTC

02/24/2009 05:17AM

192

Chapter 7

FIG. 7.6

CHART FOR CARBON AND LOW ALLOY STEELS WITH YIELD STRESS OF 30 KSI OR

GREATER, AND TYPES 405 AND 410 STAINLESS STEEL (FARR, 2006)

Example 7.3

Find the allowable compressive stress for a cylindrical shell with Ro = 25 in. and t = 0.5 in. Let T =

100F and use Fig. 7.6 for external pressure chart.

Solution

Ro/t = 50,

A = 0.125/50 = 0.0025

From Fig. 7.6, the stress value based on A = 0.0025 is in the plastic region. Hence, from Fig. 7.6, the

allowable compressive stress is B = 15,500 psi.

7.3.2 Cylindrical Shells under External Pressure and Operating at Temperatures below

the Creep Range

The ASME design equations for the lateral buckling of cylindrical shells at temperatures below the

creep range are based on Sturms work (Sturm, 1941). Sturm reduced the differential equations for the

elastic lateral buckling of a cylindrical shell to the form

Pcr = E(t/Do)3

(7.28)

where

Do = outside diameter

E = modulus of elasticity

Pcr = external buckling pressure

t = thickness

= buckling factor which is a function various parameters such as radius, thickness, length,

Poissons ratio, applied pressure, and number of buckling lobes1

Figure 7.7 (Jawad and Farr, 1989) shows a plot of for a simply supported cylindrical shell with external pressure applied on

sides and end of cylinder and a Poissons ratio of 0.3.

ASM_Jawad_Ch07.indd

192

MTC

02/24/2009 05:17AM

FIG. 7.7

COLLAPSE COEFFICIENTS OF ROUND CYLINDERS WITH PRESSURE ON SIDES AND

ENDS, EDGES SIMPLY SUPPORTED; = 0.3 (STURM, 1941)

cr = cr/E

Vcr

(7.29)

Pcr Do t

2

(7.30)

where

cr = critical circumferential buckling strain

cr = critical circumferential buckling stress

Equation (7.28) can be written as

A

ASM_Jawad_Ch07.indd

193

H cr

N

2 Do t 2

MTC

(7.31)

02/24/2009 05:17AM

194

Chapter 7

A plot of Eq. (7.31) is shown in Fig. 7.8. The (Do/t) lines in Fig. 7.8 are approximate smoothed

envelopes of those shown in Fig. 7.7. The ASME procedure consists of calculating (Do/t) and (L/Do)

values of a cylinder and then using Fig. 7.8 to obtain the critical strain A.

FIG. 7.8

GEOMETRIC CHART FOR CYLINDRICAL SHELLS UNDER EXTERNAL OR COMPRESSIVE

LOADINGS FOR ALL MATERIALS (JAWAD AND FARR, 1989)

ASM_Jawad_Ch07.indd

194

MTC

02/24/2009 05:17AM

For elastic buckling, Eqs. (7.29) and (7.30) are combined to give

Pcr

2AE

(7.32)

Do t

The ASME code uses a factor of safety of 3.0 for lateral buckling. Hence, for elastic buckling, the

governing equation for the allowable external pressure is

Pa

2AE

(7.33)

3 Do t

In the plastic region, the ASME provides pseudo stress-strain curves based on the tangent modulus

in the plastic region. This is accomplished by using a stress-strain curve for various given temperatures

plotted in terms of elastic modulus, E, and tangent modulus, Et, past the proportional limit such as

those shown in Fig. 7.6 for medium-strength carbon steels. In this case, the stress value B in Fig. 7.6

is equal to /2. The ASME procedure for calculating allowable external pressure is to determine rst

strain A from Fig. 7.8 (Eq. 7.26). Then, a value of allowable compressive stress B is determined from

Fig. 7.6. The allowable external pressure is then calculated from Eq. (7.30) as

Pcr

2 Vcr

(7.34)

Do t

Pa

4B

(7.35)

3 Do t

Example 7.4

Find the allowable external pressure, P, for a cylindrical shell with Do = 50 in., t = 0.5 in., and L =

150 in. Let T = 100F and use Fig. 7.8 for external pressure chart.

Solution

Do/t = 100,

From Fig. 7.6, it is seen that the stress value based on A = 0.00042 is in the elastic region. Hence,

either Eq. (7.33) or Eq. (7.35) may be used. From Eq. (7.33) with E = 29,000,000 psi,

P

200004229000000

3 5005

81 psi

in the Creep Range

The theoretical equations for the buckling of cylindrical and spherical shells under external pressure are extremely difcult to solve. This is because numerous variables in the equations are unknown

to the designer (Rabotnov, 1969), (Bernasconi, 1979), such as initial out-of-roundness, nal outof-roundness, time duration, and changes in modulus of elasticity and yield stress due to creep. These

variables, coupled with known variables such as length, diameter, thickness, and buckling modes,

make a closed-form solution of the buckling equations almost impossible to obtain. Accordingly,

many simplications are normally assumed by the designer. The most direct method, although fairly

approximate, is to use the material isochronous curves to simulate the effect of creep and time on buckling. These curves are available from ASME III-NH for ve materials at various creep temperatures

ASM_Jawad_Ch07.indd

195

MTC

02/24/2009 05:17AM

196

Chapter 7

ranging from 850F to 1500F. The materials are 2.25Cr-1Mo steel, 9Cr-1M0-V steel, 304 and 316

stainless steels, and nickel alloy 800H. Samples of some such isochronous curves are shown in Figs.

1.10, 2.5, 4.11, 4.13, 5.6, 6.7, and 7.9.

The isochronous curves in III-NH are not true stress-strain curves, in the theoretical sense of the

denition. They are indirectly obtained from experimental data as explained in Section 1.2.3. However, they are treated as quasi stress-strain curves in developing external pressure charts in the creep

FIG. 7.9

ISOCHRONOUS STRESS-STRAIN CURVES FOR 2.25Cr-1Mo STEEL AT 1050F

(ASME, III-NH)

ASM_Jawad_Ch07.indd

196

MTC

02/24/2009 05:17AM

range. This approximate procedure has been used for a few decades in the United States. It has been

investigated for accuracy by numerous engineers and is deemed reasonably accurate for general design

purposes. Moreover, tests performed by Dr. Don Grifn of Westinghouse show that the safety factor

can be lowered when it takes longer for the creep deformation to reach the buckling value.

The design procedure for buckling in the creep range uses the equations developed in Section 7.3 for

axial compression and external pressure in cylinders, and Section 7.4 for spherical sections. However,

the external pressure chart correlating the tangent modulus, strain A, and stress B that are given in

Fig. 7.6 for low temperatures are replaced by curves obtained from the isochronous curves for various

temperatures and times. This is illustrated by referring to the isochronous chart for 2.25Cr-1Mo steel

at 1050F shown in Fig. 7.9. It is assumed that an external pressure chart for this material is needed

for the 10,000-hour and 100,000-hour curves. This is accomplished by taking these two stress-strain

curves from Fig. 7.9 and plotting them as tangent modulus stress-strain curves in accordance with the

criterion established by ASME. The developed external pressure chart for the 2.25Cr-1Mo steel at

1050F is shown in Fig. 7.10 for 10 hours, 10,000 hours and 100,000 hours. The factors of safety are

FIG. 7.10

EXTERNAL PRESSURE CHART FOR 2.25Cr-1Mo STEEL AT 1050F

ASM_Jawad_Ch07.indd

197

MTC

02/24/2009 05:17AM

198

Chapter 7

maintained the same as those for low temperature for illustration purposes. Different factors of safety

can easily be incorporated. Similar charts can be developed for various times, temperatures, and materials as long as there are isochronous curves available.

Example 7.5

Find the allowable compressive stress for the cylindrical shell in Example 7.3. Use 10,000-hour life

from Fig. 7.10.

Solution

Ro/t = 50

A = 0.125/50 = 0.0025

From Fig. 7.10, the stress value based on A = 0.0025 is B = 2200 psi. This value is substantially

smaller than the allowable compressive stress of 15,500 psi obtained in Example 7.3.

Example 7.6

Find the allowable external pressure for the cylindrical shell in Example 7.4. Use 100,000-hour life

from Fig. 7.10.

Solution

Do/t = 100,

L/Do = 3.0,

P

4800

3100

107 psi

This allowable external pressure is more than four times smaller than that obtained at temperatures

below the creep value.

UNDER COMPRESSION

7.4.1 Spherical Shells under External Pressure and Operating at Temperatures below the

Creep Range

The theoretical equation for the elastic buckling of a spherical shell (von Karman and Tsien, 1960) is

V cr

0125 E

Ro t

(7.36)

H cr

0125

Ro t

(7.37)

or

and

ASM_Jawad_Ch07.indd

198

MTC

02/24/2009 05:17AM

Pcr

025E

Ro t 2

(7.38)

ASME uses a safety factor of 4.0 for buckling of spherical shells. Hence, the allowable external pressure in the elastic range is obtained from Eq. (7.38) as

Pa =

0.0625E

(Ro /t) 2

(7.39)

= PRo/2t

(7.40)

From Fig. 7.6 with =2B, factor of safety = 4.0, and strain A from Eq. (7.37), Eq. (7.40) becomes

Pa = B/(Ro/t)

(7.41)

Equations (7.39) and (7.41) are the two basic equations for calculating allowable external pressure

in spherical shells in accordance with the ASME code.

Example 7.7

Find the allowable external pressure, P, for a spherical shell with Ro = 25 in., t = 0.25 in. Let T =

100F and use Fig. 7.6 for external pressure chart.

Solution

Ro/t = 100,

From Fig. 7.6,

From Eq. (7.41),

A = 0.125/100 = 0.00125

B = 13,000 psi

P = 13,000/100 = 130 psi

7.4.2 Spherical Shells under External Pressure and Operating at Temperatures in the

Creep Range

The procedure for evaluating the allowable external pressure in a spherical section operating in the

creep range follows along the same lines as that for cylindrical shells. The following example serves

as an illustration.

Example 7.8

Find the allowable external pressure, P, for the spherical shell in Example 7.7. Use the 100,000-hour

curve from Fig. 7.10.

Solution

Ro/t = 100,

From Fig. 7.10,

From Eq. (7.41),

A = 0.125/100 = 0.00125

B = 1200 psi

P = 1200/100 = 12 psi

Problems

7.1 A pressure vessel is supported by 316 stainless steel legs that are 12 in. std wt pipe, 4 ft long.

The design temperature of the legs is 1200F and the load on each leg is 5000 lb. Check the

adequacy of the legs for an expected life of 100,000 hours. Let

n = 4.0

K = 2.45 10-18

ASM_Jawad_Ch07.indd

199

MTC

02/24/2009 05:17AM

200

Chapter 7

r = 4.38 in.

I = 279 in.4

A = 14.6 in.2

FS = 1.2

Smt = 5500 psi at 100,000 hours

7.2 A pressure vessel is operating at 1050F. The outside diameter is 12 ft, shell thickness is 4.0

in., and the effective length is 24 ft. Check the shell for external pressure due to a vacuum

condition of -15 psig using Fig. 7.10 and 100,000 hours.

ASM_Jawad_Ch07.indd

200

MTC

02/24/2009 05:17AM

APPENDIX A

BACKGROUND OF THE

BREE DIAGRAM

A.1

The Bree diagram (Fig. A.1) is constructed for the purpose of combining mechanical and thermal

stresses in a cylindrical shell in order to establish allowable stress criteria. The diagram is plotted with

the mechanical stress as an abscissa and the thermal stress as an ordinate. Zones E, S1, S2, P, R1, and

R2, shown in the gure, correspond to specic mechanical and thermal load groupings. The criteria

used to establish each of these zones are described below and are based on Bree (1967 and 1968),

Burgreen (1975), Kraus (1980), and Wilshire (1983).

Zone E

It is assumed that all stresses in zone E remain elastic in the rst half (thermal loading) as well as

the second half (thermal down loading) of the thermal cycle. Bree assumed a thin cylindrical shell

subjected to internal pressure. The average stresses in the cylinder are

= PRi /t

(A.1)

l = PRi / 2t

(A.2)

(A.3)

where

P

Ri

t

l

r

= internal pressure

= inside radius

= thickness of cylinder

= longitudinal stress

= radial stress

= circumferential stress

Radial thermal gradients in thin-walled cylinder are normally linear in distribution. For a linear

radial thermal distribution through the thickness (Fig. A.2), the thermal stress equations (Burgreen,

1975) are expressed as

= l = (2x/t){EDT / [2(1 - )]}

(A.4)

201

ASM_Jawad_BM_AppendixA.indd

201

MTC

02/24/2009 04:30AM

202

Appendix A

FIG. A.1

BREE DIAGRAM (BREE, 1967)

r 0

(A.5)

where

E = modulus of elasticity

x = distance from midwall of the cylinder (Fig. A.2)

= coefcient of thermal expansion

T = difference in temperature between inside and outside surfaces of the cylinder

= Poissons ratio

ASM_Jawad_BM_AppendixA.indd

202

MTC

02/24/2009 04:30AM

FIG. A.2

THERMAL GRADIENT IN SHELL

Bree made the following assumptions in order to evaluate the mechanical and thermal stress equations in a practical manner:

Radial mechanical and thermal stresses, r, are small compared to the circumferential and

longitudinal stresses and can thus be ignored.

Because mechanical stress is considered primary stress, it cannot exceed the yield stress value

of the material. Radial thermal stress, on the other hand, is considered secondary stress and

can thus exceed the yield stress of the material.

The combination of mechanical and thermal stresses in the and l directions may result in

stresses past the yield stress in one direction and not the other. Such a condition prevents the

formulation of a closed-form solution in a cylindrical shell. Accordingly, Bree made a conservative assumption by setting the stress in the l direction, l, equal to zero.

With the l and r stresses set to zero, Bree assumed for simplicity that the remaining stress in the

direction could be applied on a at plate as shown in Fig. A.3 with values obtained from

Eqs. (A.1) and (A.4) for cylindrical shells.

An additional assumption was made where the at plate is not allowed to rotate due to variable

thermal stress across the thickness in order to simulate the actual condition in a cylindrical

shell.

FIG. A.3

STRESS IN THE CIRCUMFERENTIAL DIRECTION

ASM_Jawad_BM_AppendixA.indd

203

MTC

02/24/2009 04:30AM

204

Appendix A

It is assumed that the material has an elastic perfectly plastic stress-strain diagram.

The initial evaluation of the mechanical and thermal stresses in the elastic and plastic region

was made without any consideration to relaxation or creep.

The nal results were subsequently evaluated for relaxation and creep effect.

It is assumed that stress due to pressure is held constant, whereas thermal stress is cycled.

Hence, pressure and temperature stress exist at the end of the rst half of the cycle, and only

pressure exists at the end of the second half of the cycle.

Based on these assumptions, the mechanical and thermal stress in zone E can now be derived. It

is assumed that the stresses remain elastic during the stress cycle. The stress distribution in the rst

half-cycle, shown in Fig. A.4, is

= PRi /t + (2z/t){EDT /[2(1 - )]}

(A.6)

= p + (2z/t)T

(A.7)

or

where

z = distance from midwall of the at plate (Fig. A.2)

p = stress due to pressure, PRi/t

T = stress due to temperature, EDT/2(1 - )

It should be noted that for a at plate, Eq. (A.6) uses the quantity DT/[2(1 - )] rather than DT in

order to simulate thermal stress in a cylinder.

The maximum value of Eq. (A.7) is reached when z = t/2. In zone E, the stress combination of p

plus T is kept below y, as shown in Fig. A.4. Thus, Eq. (A.7) becomes

p + T < y

(A.8a)

Dene

X = p /y

and

Y = T/y

FIG. A.4

STRESS CYCLE IN ZONE E

ASM_Jawad_BM_AppendixA.indd

204

MTC

02/24/2009 04:30AM

X + Y = 1.0

(A.8b)

p < y

(A.9)

Thus, zone E, which denes an elastic stress condition, is bound by Eq. (A.8) as well as the x- and

y-axes as shown in Fig. A.1.

Zone S1

In zone S1, it is assumed that shakedown will occur after the rst cycle. It is also assumed that the

elastic pressure plus temperature stress combination in the rst half-cycle exceeds the yield stress of

the material at only one side of the shell, as shown in Fig. A.5. Stress in the second half-cycle after

the temperature is removed remains elastic. This assures shakedown after the rst cycle. The strain

distribution in the elastic region is

E1 = 1 - (2z/t)T

(A.10)

where the subscript 1 refers to the rst half-cycle and v is to be determined. The strain distribution

in the plastic region is

E1 = y - (2z/t)T + E

(A.11)

At point v, 1 = y. Equating Eqs. (A.10) and (A.11) gives

E = 2[(z - v)/t]T

(A.12)

1 = y + 2[(z - v)/t]T

(A.13)

1 = y

(A.14)

The expression for v is obtained by summing the stress Eqs. (A.13) and (A.14) across the thickness

in accordance with the equation

t/2

- t/2

dz = tp

(A.15)

FIG. A.5

STRESS CYCLE IN ZONE S1 WITH YIELDING ON ONE SIDE ONLY

ASM_Jawad_BM_AppendixA.indd

205

MTC

02/24/2009 04:30AM

206

Appendix A

This gives

v/t = [(y - p)/T]1/2 - 1/2

(A.16)

A limit of Eq. (A.16) is that v cannot be less than zero. This gives

p + T/4 < y

(A.17a)

X + Y/4 = 1.0

(A.17b)

Another limit of Eq. (A.16) is that v must be less than t/2. This gives

p + T < y

(A.18a)

X + Y = 1.0

(A.18b)

or

In order for shakedown to take place, the stresses remaining after removal of the temperature in

the second half-cycle must remain elastic. The change of strain from the rst to second half-cycle is

given by

ED2 = D2 + (2z/t)T

= 2 - 1 + (2z/t)T

(A.19)

(A.20)

and

ED2 = D2 + (2z/t)T

= 2 - y + (2z/t)T

(A.21)

(A.22)

where the subscript 2 refers to the second half-cycle. Multiplying both sides of Eqs. (A.19) and

(A.21) by dz and integrating their total sum results in D2 = 0. This is because the integral of the total

sum of D2 = 0 since there is no change in external forces during the second half-cycle. Also, the total

sum of the integral (2z/t)T from the limits -t/2 to +t/2 is equal to zero. Hence,

D2 = 0

(A.23)

Substituting this quantity and Eq. (A.13) into Eqs. (A.20) and (A.22) results in

2 = y - (2v/t)T

(A.24)

2 = y - (2z/t)T

(A.25)

One of the conditions for shakedown is that 2 < y. Hence, the above two equations reduce to

v>0

(A.26)

T < 2y

(A.27a)

Y < 2.0

(A.27b)

and

or

Thus, zone S1 is bounded by Eqs. (A.17), (A.18), and (A.27). This is shown by area BCDF in Fig.

A.1.

ASM_Jawad_BM_AppendixA.indd

206

MTC

02/24/2009 04:30AM

Zone S2

In zone S2, it is assumed that shakedown will occur after the rst cycle. It is also assumed that the

elastic pressure plus temperature stress combination in the rst half-cycle exceeds the yield stress of

the material at both sides of the shell. Stress in the second half-cycle after the temperature is removed

remains elastic. This assures shakedown after the rst cycle. The derivation of the limiting equations is

similar to that in zone S1 with the exception that both sides of the shell reach the yield stress, as shown in

Fig. A.6. Two unknown quantities v and w must be determined. The solution (Burgreen, 1975) yields

v/t = (1/2)[(y/T) - (p/y)]

(A.28)

(A.29)

A limiting value of these two equations is v < t/2 and w < -t/2. The two equations reduce to

T(y + p) = y2

(A.30)

T (y - p) = y2

(A.31a)

Y(1 - X) = 1.0

(A.31b)

and

or

The second of these equations is the prevalent one, because satisfying it will automatically satisfy

the rst one. Equation (A.27) is also a controlling equation in the S2 domain. Thus, Eqs. (A.27) and

(A.31) form the boundary of zone S2. This is shown by area CDF in Fig. A.1. Notice that this zone falls

within zone S1. Thus, zone S2, where yielding occurs on both sides of the shell, supersedes the condition in zone S1 in the range shown.

Zone P

Plasticity is assumed to occur in zone P. The main characteristics of this zone is that the core section of the shell remains elastic (otherwise, ratcheting may occur) whereas the outer surfaces alternate

between tensile and compression yield stress as the temperature is applied then removed as shown in

Fig. A.7. Figure A.7a shows the stress distribution at the end of the rst half-cycle. It resembles the

same gure in the second half-cycle of zone S2. Figure A.7b shows the stress distribution during the

second half-cycle, whereas Fig. A.7c shows the stress distribution at the end of the second half-cycle

FIG. A.6

STRESS CYCLE IN ZONE S2 WITH YIELDING ON BOTH SIDES

ASM_Jawad_BM_AppendixA.indd

207

MTC

02/24/2009 04:30AM

208

Appendix A

FIG. A.7

STRESS CYCLE IN ZONE P

(Burgreen, 1975). Referring to Fig. A.7 with the center portion of the shell remaining elastic, -u z u,

the relationship between the stress and strain (Kraus, 1980) is given by

ED2 = D2 + (2z/t)T

(A.32)

these two values in Eq. (A.32) and subtracting the resultant two equations, since the strain change is

linear, gives

u = (y/T)t

(A.33)

T 2y

(A.34a)

Y2

(A.34b)

or,

The net stress distribution in the core section of the cylinder (Fig. A.7c) is

= y - 2T(v/t)

-w z v

(A.35)

This stress cannot exceed the yield stress. Thus, Eq. (A.35) gives v 0. Substituting this value in

Eq. (A.28) gives

pT y2

(A.36a)

(X )(Y ) = 1

(A.36b)

or

Equations (A.34) and (A.36) form the boundary of zone P. It is dened by area DFG of the Bree

diagram in Fig. A.1.

ASM_Jawad_BM_AppendixA.indd

208

MTC

02/24/2009 04:30AM

Zone R1

Ratcheting is assumed to take place in zone R1. The main characteristic of this zone is that yielding

extends past the midwall of the shell due to yielding at one of the surfaces of the shell. Thus, shakedown will not take place. Figure A.8 shows the stress distribution in the rst and second half-cycles,

whereas Fig. A.9 shows the stress distribution in the second and third half-cycles.

The rst half-cycle (Fig. A.8a) is the same as the stress distribution in zone S1. The stress-strain

relationship for the second half-cycle (Fig. A.8c) is given by

ED2 = y - 1 + (2z/t)T + ED2

-t/2 z v

(A.37)

v z v

(A.38)

ED2 = 2 - y + (2z/t)T

v z t/2

(A.39)

From Eqs. (A.37), (A.38), and (A.39) plus the boundary conditions shown in Fig. A.8c, the following equations are obtained after various substitutions:

2 = y + 2T(v - z)/t

v z t/2

(A.40)

2 = y

-t/2 z v

(A.41)

1/2

(A.42)

ED2 = 4Tv /t

-t/2 z v

(A.43)

ED2 = 2T (v - z)/t

v z v

(A.44)

In the third half-cycle (Fig. A.9), the temperature stress is re-applied. The stress-strain relationship

for the third half-cycle (Fig. A.9c) is given by

ED3 = 3 - y - (2z/t)T

-t/2 z v

(A.45)

v z v

(A.46)

v z t/2

(A.47)

FIG. A.8

FIRST AND SECOND HALF CYCLES IN ZONE R1

ASM_Jawad_BM_AppendixA.indd

209

MTC

02/24/2009 04:30AM

210

Appendix A

FIG. A.9

SECOND AND THIRD HALF CYCLES IN ZONE R1

From Eqs. (A.45), (A.46), and (A.47) plus the boundary conditions shown in Fig. A.9c, the following equations are obtained after various substitutions:

3 = y + 2T(z - v)/t

-t/2 z v

(A.48)

3 = y

v z t/2

(A.49)

(A.50)

v z v

(A.51)

v z t/2

(A.52)

The plastic strain through the full cycle is obtained by adding Eqs. (A.44) and (A.52) to obtain

ED = 4 Tv/t

(A.53)

ED = 4T{0.5 - [(y - p)/T)]1/2}

(A.54)

(p + T/4) y

(A.55a)

X + Y/4 1.0

(A.55b)

or

The second requirement is that 2 in Eq. (A.40) must be greater than -y at z = t/2. Similarly, 3 in

Eq. (A.48) must be greater than -y at z = -t/2. Using one of these two equations yields

T(y - p) y2

(A.56a)

Y(1 - X) 1.0

(A.56b)

or

ASM_Jawad_BM_AppendixA.indd

210

MTC

02/24/2009 04:30AM

The nal requirement is that the stress due to pressure, p, must be less than the yield stress

p y

(A.57a)

X 1.0

(A.57b)

or

Equations (A.55), (A.56), and (A.57) form the boundary of the zone R1. It is dened by area BFH

of the Bree diagram in Fig. A.1.

Zone R2

Ratcheting is assumed to take place in zone R2. The main characteristic of this zone is that yielding

extends past the midwall of the shell due to yielding at both surfaces of the shell. Thus, shakedown will

not take place. The derivation of the equations is very similar to that for zone R1. The area for zone R2

is dened by area FGH of the Bree diagram in Fig. A.1.

ASM_Jawad_BM_AppendixA.indd

211

MTC

02/24/2009 04:30AM

APPENDIX B

CONVERSION TABLE

Multiply

By factor

To get units

bar

cm

inches

kg

kg/mm2

lb

MPa

mm

psi

psi

psi

C

F

14.504

0.3937

25.4

2.205

14.22

0.454

145.03

0.03937

0.06895

0.0704

0.006895

1.8C + 32

(F - 32)/1.8

psi

inches

mm

lb

psi

kg

psi

inches

bar

kg/mm2

MPa

F

C

212

ASM_Jawad_BM_AppendixB.indd

212

MTC

02/19/2009 12:13PM

REFERENCES

American Institute of Steel Construction, 1991, Manual of Steel Construction-Allowable Stress

Design, ASCE, New York.

American Petroleum Institute, 2000, Fitness-for-Service: API Recommended Practice 579, API,

Washington, D.C.

American Society of Mechanical Engineers, 1976, ASME-MPC symposium on Creep-Fatigue Interaction, ASME, New York.

American Society of Mechanical Engineers, 2007, Boiler and Pressure Vessel Code, Section I, Rules

for Construction of Power Boilers, ASME, New York.

American Society of Mechanical Engineers, 2007, Boiler and Pressure Vessel Code, Section II, Materials, ASME, New York.

American Society of Mechanical Engineers, 2007, Boiler and Pressure Vessel Code, Section III, Rules

for Construction of Nuclear Facility Components, ASME, New York.

American Society of Mechanical Engineers, 2007, Boiler and Pressure Vessel Code, Section VIII,

Rules for Construction of Pressure Vessels, ASME, New York.

ASME, 2004, American Society of Mechanical Engineers, Power Piping ASME B31.1, New York,

ASME.

ASME, 2004, American Society of Mechanical Engineers, Power Piping ASME B31.3, New York,

ASME.

Baker, J., and Heyman, J., 1969, Plastic Design of Frames, University of Cambridge Press, Cambridge.

Bannantine, J. A., Comer, J. J., and Handrock, J. L., 1990, Fundamentals of Metal Fatigue Analysis,

Prentice-Hall, New Jersey.

Becht, C., et al., 1989, Structural Design for Elevated Temperature Environments Creep,

Ratchet, Fatigue, and Fracture, PVP-163, NY, American Society of Mechanical Engineers.

Beedle, L., 1958, Plastic Design of Steel Frames, John Wiley Publishing, New York.

Bernasconi, G., and Piatti, G., 1979, Creep of Engineering Materials and Structures, Applied Science

Publishers Ltd., London.

Boyle, J. T., and Spence, J., 1983, Stress Analysis for Creep, Camelot Press, London.

Bree, J., 1967, Elastic-Plastic Behaviour of Thin Tubes Subjected to Internal Pressure and Intermittent High-Heat Fluxes with Application to Fast-Nuclear-Reactor Fuel Elements, Journal of Strain

Analysis, 2(3), pp. 226238.

Bree, J., 1968, Incremental Growth Due to Creep and Plastic Yielding of Thin Tubes Subjected to

Internal Pressure and Cyclic Thermal Stresses, Journal of Strain Analysis, 3(2).

213

ASME_Jawad_Reference.indd

213

MTC

02/24/2009 06:37AM

214 References

Burgreen, D., 1975, Design Methods for Power Plant Structures, C. P. Press, New York.

Chen, W. F., and Zheng, H., 1991, Structural Plasticity, Springer-Verlag, New York.

Conway, J. B., 1969, Stress-Rupture Parameters, Origin, Calculations and Use, Gordon and Breach

Science Publishers, New York.

Curran, R. M., 1976, Symposium on Creep Fatigue Interaction, MPC-3, American Society of

Mechanical Engineers, New York.

Den Hartog, J. P., 1987, Advance Strength of Materials, Dover Publications, Massachusetts.

Dillon, C. P., 2000, Unusual Corrosion Problems in the Chemical Industry, Materials Technology Institute of the Chemical Process Industries, Missouri.

Farr, J. R., and Jawad, M. H., 2006, Guidebook for the Design of ASME Section VIII Pressure Vessels,

ASME Press, New York.

Faupel, J. H., and Fisher, F. E., 1981, Engineering Design, John Wiley, New York.

Finnie, I., and Heller, W., 1959, Creep of Engineering Materials, McGraw Hill, New York.

Flugge, W., 1967, Viscoelasticity, Blaisdell Publishing Company, Massachusetts.

Frost, N. E., Marsh, K. J., and Pook, L. P., 1974, Metal Fatigue, Clarendon Press, Oxford.

Goodall, I. W., 2003, Reference Stress Methods Analysing Safety and Design, Professional Engineering Publishing, London.

Grant, N. J., and Mullendore, A. W., 1965, Deformation and Fracture at Elevated Temperatures, MIT

Press, Massachusetts.

Hill, R., 1950, The Mathematical Theory of Plasticity, Oxford Press, London.

Hoff, N. J., 1958, High Temperature Effects in Aircraft Structures, Pergamon, New York.

Hult, J., 1966, Creep in Engineering Materials, Blaisdell Publishing Company, Massachusetts.

Hyde, T. H., 1994, Creep of Materials and Structures, Mechanical Engineering Publications, London.

Jawad, M. H., 2004, Design of Plate and Shell Structures, ASME Press, New York.

Jawad, M. H., and Farr, J. R., 1989, Structural Analysis and Design of Process Equipment, John Wiley,

New York.

Jetter, R. I., 2002, Companion Guide to the ASME Boiler & Pressure Vessel Code, Rao, K. R., ed.,

ASME Press, New York, Volume 1, Chapter 12.

Kraus, H., 1980, Creep Analysis, John Wiley Publishing, New York.

Larsson, L. H., 1992, High Temperature Structural Design, ESIS 12, Mechanical Engineering Publications, London.

Markl, A., 1960, Fatigue Tests of Piping Components, Pressure Vessel and Piping Design, Collected

Papers, 19271959, ASME, New York.

Neuber, H., 1961, Theory of Stress Concentration for Shear-Strained Prismatical Bodies With

Arbitrary Nonlinear Stress-Strain Law, ASME Journal of Applied Mechanics, 28, pp. 544550.

ODonnell, W. J., and Porowski, J., 1974, Upper Bounds for Accumulated Strains Due to Creep

Ratcheting, Transactions ASME Journal of Pressure Vessel Technology, 96, pp. 150154.

ASME_Jawad_Reference.indd

214

MTC

02/24/2009 06:37AM

References 215

Odqvist, F. K. G., 1966, Mathematical Theory of Creep and Creep Rupture, Oxford Mathematical

Monographs, Oxford.

Penny, R. K., and Marriott, D. L., 1995, Design for Creep, Chapman and Hall, New York.

Prager, M., 1995, Development of the MPC Omega Method for Life Assessment in the Creep

Range, Journal of Pressure Vessel Technology, 117, pp. 95103.

Prager, M., 2000, The Omega Method - An Engineering Approach to Life Assessment, Journal of

Pressure Vessel Technology, Vol. 122, American Society of Mechanical Engineers, New York.

Rabotnov, Yu. N., 1969, Creep Problems in Structural Members, Elsevier, New York.

Rao, K. R., 2002, Companion Guide to the ASME Boiler and Pressure Vessel Code, ASME, New York.

Severud, L. K., 1975, Simplied Methods and Application to Preliminary Design of Piping for

Elevated Temperature Service, ASME 2nd National congress on pressure vessels and piping, San

Francisco, ASME.

Severud, L. K., 1980, Experience With Simplied Inelastic Analysis of Piping Designed for

Elevated Temperature Services, ASME Journal of Nuclear Engineering, ASME, New York.

Severud, L. K., 1991, Creep-Fatigue Assessment Methods Using Elastic Analysis Results and

Adjustments, ASME Journal of Pressure Vessel Technology, 114, pp. 3440.

Severud, L. K., and Winkel, B. V., 1987, Elastic Creep-Fatigue Evaluation for ASME Code,

Transactions of the 9th International conference on structural mechanics in reactor technology,

Lausanne, Switzerland.

Sim, R. G., 1968, Creep of structures, Ph.D. dissertation, University of Cambridge, Cambridge,

U.K.

Smith, A. I., and Nicolson, A. M., 1971, Advances in Creep Design, John Wiley, New York.

Sturm, R. G., 1941, A Study of the Collapsing Pressure of Thin-Walled Cylinders, University of

Illinois Engineering Experiment Station Bulletin 329, Urbana, Illinois.

Thielsch, H., 1977, Defect and Failures in Pressure Vessels and Piping, Robert E. Krieger Publishing,

New York.

Von Karman, T., and Tsien, H., 1960, The Buckling of Spherical Shells by External Pressure,

Pressure Vessel and Piping Design Collected Papers 19271959, ASME, New York.

Wang, C. K., 1970, Matrix Methods of Structural Analysis, International Textbook Company, Pennsylvania.

Weaver, W., and Gere, J., 1990, Matrix Structural Analysis, Van Nostrand, New York.

Wilshire, B., and Owen, D. R. J., 1983, Engineering Approaches to High Temperature Design, Pineridge

Press, Swansea, U.K.

Winston, B., Burrows, R., Michel, R., and Rankin, A.W., 1954. A Wall-Thickness Formula for

High-Pressure Piping, Transactions of ASME, New York.

ASME_Jawad_Reference.indd

215

MTC

02/24/2009 06:37AM

ASME_Jawad_Reference.indd

216

MTC

02/24/2009 06:37AM

- 05_Process Piping - The Complete Guide to ASME B31.3 - Becht 3rd EdUploaded byKIM
- Guide to ASME Code VIII, Division 1, Pressure VesselsUploaded byStefano Milani
- En13445 vs AsmeUploaded byjeff501
- PTB-2_Edtn_2009.pdfUploaded byHenry Javier Ríos
- PTB-7-2014.pdfUploaded byLuisMedina
- Welding Research Council Bullentin - WRC 537Uploaded byjontylee87
- Power Piping-The Complete Guide to Asme b31.1 - 2013 - By Charles Becht IVUploaded byRusbel Rios
- 35057Uploaded byAl Nanbai Caspian
- Pressure VesselUploaded byAlejandro Dávila
- Asme CourseUploaded bySajal Kulshrestha
- ASME_PTB-5-2013_cont_0101Uploaded bypouya_ms
- Guide to ASME Sec IUploaded bynagarathinam82
- ASME VIII Div 1 Fact BookUploaded byARNOUX
- Companion Guide to the ASME Boiler & Pressure Vessel Code Volume 1Uploaded bysusanweb
- Structural Analysis and Design of Process EquipmentUploaded bySurya Tej
- Asme Viii 2015Uploaded byoilandgasoline
- Process Piping_The Complete Guide to ASME B31.3 -Charles Becht IVUploaded byalan172081
- ASME III Design by Analysis - 2006Uploaded byTimothy Hull
- ASME Pressure Vessels Basic CalculationsUploaded bydhasdj
- ASME VIII Div. 1- 1 Internal PressureUploaded byعصمت ضيف الله
- Hand Book Asme 9Uploaded byTarek Baccari
- Guidebook ASME Section VIII.pdfUploaded bybensossi_441944896
- Pressure Vessel Design - Guides and ProceduresUploaded bymatteo_1234
- Design Analysis of PVUploaded byDivakar Panigrahi
- Theory and Design of Pressure Vessels - JOHN F. HARVEY, P.EUploaded byRajwinder Singh
- Secrets to Using ASME Section IX Welding and BrazingUploaded byMichelle Mc Intyre
- Asme b31.3 Training SeminarUploaded byglazetm
- ASME Code CalculationsUploaded byPunith Rangegowda

- Dialnet-MedicionDeTemperaturaSensoresTermoelectricos-4806937Uploaded byDaniel Andres Doza Chavita
- drone con arduino.pdfUploaded byOsmar Gabriel Peñaloza
- Ecodial_4.8_user_manual.pdfUploaded byErnesto Mora
- 315033002 Digital Signal Processing Principles Algorothims and Applications 4th Edition John G ProakisUploaded byOsmar Gabriel Peñaloza
- MPLAB XC8 C Compiler User GuideUploaded byOsmar Gabriel Peñaloza
- Temario Galpón con armadura (LRDF) 1-2017 (1).pdfUploaded byOsmar Gabriel Peñaloza
- Temario Galpón con armadura (LRDF) 1-2017 (1).pdfUploaded byOsmar Gabriel Peñaloza
- CINEMÁTICAUploaded byOsmar Gabriel Peñaloza
- Los jóvenes y el uso de las drogas en la España de los 90 tesis 1994Uploaded byAlberto Domínguez Martínez
- (Carlos Pardo) - Flujo CompresibleUploaded byOsmar Gabriel Peñaloza
- MQ-2Uploaded byapi-3850017
- MPLAB C18 Libraries 51297fUploaded byGeorgy Parada Mass
- 07.pdfUploaded byOsmar Gabriel Peñaloza
- Clase 13.2 - Comunicaciones IndustrialesUploaded byOsmar Gabriel Peñaloza
- Links LibrosUploaded byRobinson Chiran Acosta
- Una Sencilla Regla Para Predimensionar Pilares de HormigónUploaded byjowacoco
- quemadores de gas con premescla de aire arrastrado por el chorroUploaded byMarcelo Braconi
- Fl_sim_hUploaded byAngodi Maniz
- structural wood elements design manualUploaded byje_free_666
- SÍMBOLOUploaded byOsmar Gabriel Peñaloza
- DISEÑO Y DIMENSIONAMIENTO DE BOMBAS.docxUploaded byOsmar Gabriel Peñaloza
- Tomo1CaratulaUploaded byNestor Orozco
- modelado_matematico.pdfUploaded byEriiqa Serrano
- Hie MesaUploaded bylucas.barravaite
- Machine Learning y SensoresUploaded byOsmar Gabriel Peñaloza
- Minicurso en c Xc8Uploaded byOsmar Gabriel Peñaloza
- Art RitisUploaded byOsmar Gabriel Peñaloza
- 2 Motores lubricacionUploaded byErick Javier Torres Manotas.
- Apuntes de Control PidUploaded bypaulmejiaglz
- Norma Din AceroUploaded byJorge Ignacio Mora Silva

- Stresses at Junction of Cone and Cylinder In Tanks With Cone Bottoms or Ends - BoardmanUploaded byhawkwhite
- Lupranol_3402Uploaded byA Mahmood
- Theory and Application of RF&Microwave AbsorbersUploaded byBhargav Bikkani
- chapter_8Uploaded bySunil kumar Kumawat
- ANSYS FLUIDSUploaded bycapitan55
- HES2125_ Lec11&12Uploaded byultdex
- 10 Years of Dual NatureUploaded bydipeshjoon
- DNV-OS-C501 - 2013.pdfUploaded bymecanicaufsj2013
- Commissioning Procedure of Northern Gas Compression FacilitiesUploaded byMohamed Sayed Abdo
- Hysteresis of Bed Forms During Changing StageUploaded byAaliyan Syed
- LRB Kontrol pembebananUploaded byIndrayonManalu
- Composite Materials DesignUploaded byGihanFernando
- 31028885Uploaded byLim Teck Huat
- Plate TutorialUploaded byKhaled Chaib
- INSTAFLEX technical manual 2006.pdfUploaded byCristianDumitru
- Torque ConvertersUploaded bychungfung
- hardness-preheat.pdfUploaded bydanghp
- Entanglement The Greatest Mystery in Physics By Amir D. Aczel (Abee).pdfUploaded byhadzimusic
- A Review of Intensification of Photocatalytic ProcessesUploaded byJackson Ilha
- Shirs AtUploaded byDrPrasanta Ghosh
- 067_Antifrication Bearing Maintenance & LubeUploaded bynazir305
- 1016 VEM2003-20577-C14-01 ESPINOUploaded byaptureinc
- Piping Tie-Rod Design Made SimpleUploaded bySeungmin Paek
- New York Tsunami Mitigation PlanUploaded byStephen Loiaconi
- AcousticsUploaded bySudhir Swarnkar
- EPS ASME IX e ASME B31.3 Gtaw+Smaw Aço API 5L Gr.B.pdfUploaded byLeandro Sabino
- SPE-59333-MSUploaded bywsanchez_sotelo
- MEMS Paper Ppt[1]Uploaded byGaurav Jodh
- Binocolo Edf Description UseUploaded byMicheleFiorentino
- Chapter 10 LectureUploaded byMeshal Al-mutairi